University of Southern California Dissertations and Theses

Roadway hazard analysis: a safe ride for motorcycles

(USC Thesis Other)

Roadway hazard analysis: a safe ride for motorcycles

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content

Roadway Hazard Analysis:
A Safe Ride for Motorcycles

by

Emily Carol Bartee

A Thesis Presented to the
Faculty of the USC Dornsife College of Letters, Arts and Sciences
University of Southern California
In Partial Fulfillment of the
Requirements for the Degree
Master of Science
(Geographic Information Science and Technology)

May 2020

Copyright © 2020 by Emily Bartee

To those who have lost their lives in the data used for this thesis – you are more than a statistic.

iv
Table of Contents
List of Figures ............................................................................................................................... vii
List of Tables ................................................................................................................................. ix
Acknowledgments........................................................................................................................... x
List of Abbreviations ..................................................................................................................... xi
Abstract ......................................................................................................................................... xii
Chapter 1 Introduction .................................................................................................................... 1
1.1. Motivation ...........................................................................................................................3
1.1.1. Natural Vulnerability .................................................................................................3
1.1.2. Statistics and Trends ..................................................................................................5
1.1.3. Existing Hazard Analysis ...........................................................................................5
1.2. Study Area ..........................................................................................................................7
1.2.1. The Commonwealth of Kentucky ..............................................................................8
1.2.2. Livingston County, Kentucky ....................................................................................9
1.2.3. Meade County, Kentucky ........................................................................................11
1.2.4. Jefferson County, Kentucky.....................................................................................12
1.2.5. Fayette County, Kentucky .......................................................................................14
1.3. Thesis Organization ..........................................................................................................15
Chapter 2 Related Work................................................................................................................ 17
2.1. Identified Factors ..............................................................................................................17
2.2. Motorcycle Specific Mobile Applications ........................................................................19
2.3. Current Safety Initiatives ..................................................................................................20
2.3.1. Behavior-Based Initiatives .......................................................................................21
2.3.2. Infrastructure-Based Initiatives ................................................................................22
Chapter 3 Data and Methods......................................................................................................... 24
v
3.1. Data Description ...............................................................................................................25
3.1.1. Jurisdiction and Base Layers ...................................................................................25
3.1.2. Collision Factor Data Layer .....................................................................................26
3.1.3. Recorded Crash Layer..............................................................................................27
3.2. Research Design................................................................................................................28
3.2.1. Configure Layers and Map Document .....................................................................29
3.2.2. Utilize Optimized Hot Spot Analysis Tool ..............................................................33
3.2.3. Identify Suitable Roadway Segments ......................................................................35
3.2.4. Summarize Viable Motorcycle Lane Locations ......................................................39
Chapter 4 Results .......................................................................................................................... 40
4.1. Livingston County Results ................................................................................................40
4.2. Meade County Results ......................................................................................................42
4.3. Jefferson County Results ..................................................................................................43
4.4. Fayette County Results .....................................................................................................45
Chapter 5 Conclusion .................................................................................................................... 48
5.1. Project Limitations ............................................................................................................49
5.2. Future Work ......................................................................................................................52
References ..................................................................................................................................... 55
Appendix A: Queries .................................................................................................................... 58
Appendix B: Results ..................................................................................................................... 59
Appendix C: Final Results ............................................................................................................ 60
Appendix D: Commonwealth of Kentucky .................................................................................. 61
Appendix E: Livingston County ................................................................................................... 65
Appendix F: Meade County .......................................................................................................... 68
Appendix G: Jefferson County ..................................................................................................... 72
vi
Appendix H: Fayette County ........................................................................................................ 80

vii
List of Figures
Figure 1. Motorcycle comparison ................................................................................................... 4
Figure 2. Commonwealth of Kentucky ........................................................................................... 7
Figure 3. Motorcycle representation ............................................................................................... 8
Figure 4. Kentucky crash fatalities ................................................................................................. 9
Figure 5. Highest county injury percentage .................................................................................. 10
Figure 6. Map of Livingston County ............................................................................................ 10
Figure 7. Highest county fatality percentage ................................................................................ 11
Figure 8. Meade County ............................................................................................................... 12
Figure 9. Highest collision count by county ................................................................................. 13
Figure 10. Highest injury count by county ................................................................................... 13
Figure 11. Highest fatality count by county .................................................................................. 13
Figure 12. Jefferson County .......................................................................................................... 14
Figure 13. Fayette County............................................................................................................. 15
Figure 14. Methods outline ........................................................................................................... 29
Figure 15. Layers and symbology ................................................................................................. 30
Figure 16. OHSA example ............................................................................................................ 35
Figure 17. Clip layer tool .............................................................................................................. 38
Figure 18. Clip for AADT ............................................................................................................ 38
Figure 19. Clip for widening ......................................................................................................... 38
Figure 20. Summary statistics ....................................................................................................... 39
Figure 21. Livingston County final analysis ................................................................................. 41
Figure 22. Meade County final analysis ....................................................................................... 43
Figure 23. Jefferson final analysis ................................................................................................ 44
Figure 24. Fayette County final analysis ...................................................................................... 46
viii
Figure 25. Jefferson County 250 m cell ........................................................................................ 51
Figure 26. Jefferson County default cell ....................................................................................... 51
Figure 27. Meade County example ............................................................................................... 52

ix
List of Tables
Table 1 Jurisdiction and base layers ............................................................................................. 26
Table 2 Collision factor layers ...................................................................................................... 27
Table 3 Recorded collision data .................................................................................................... 27
Table 4 Livingston County results ................................................................................................ 41
Table 5 Meade County results ...................................................................................................... 42
Table 6 Jefferson County results................................................................................................... 45
Table 7 Fayette County results ..................................................................................................... 47

x
Acknowledgments
The gratitude I have for the help in completing this thesis is immeasurable. First, to my thesis
advisor Dr. Bernstein: thank you for the continuous guidance, direction, and motivation to make
this document complete and purposeful. I traveled down many rabbit holes and your experience
and insight kept me on track to a successful and meaningful thesis. Second, to my coworkers at
the Kentucky Transportation Cabinet: thank you for the hints, institutional knowledge, and
constant support in completing my Masters. Your willingness to help me grow as a professional
and academic will not be taken for granted. Finally, to my boyfriend and family: thank you for
being patient with me as I committed my time to the research and technical work for this project.
From making sure I could work uninterrupted, reminding me to take care of myself, and listening
to me lament on the many hurdles experienced, I cannot express enough how thankful I am.

xi
List of Abbreviations
AADT Average Annual Daily Traffic
AGOL ArcGIS Online
CU Horizontal Curve
DGI Department of Geographic Information
Esri Environmental Systems Research Institute
EV Rating Evaluation Section
GIS Geographic information system
GISci Geographic information science
GR Grade or Vertical Curve
HLDI Highway Loss Data Institute
IIHS Insurance Institute for Highway Safety
KSP Kentucky State Police
KYTC Kentucky Transportation Cabinet
LN Lane Width
OHSA Optimized Hot Spot Analysis
NAD North American Datum
NHSTA National Highway Traffic Safety Administration
SL Speed Limit
SSI Spatial Sciences Institute
TF Traffic Flow
USC University of Southern California

xii
Abstract
Motorcycles are disproportionately affected in collisions when compared to other motor
vehicle types, leading to an increased vulnerability of injury or death to motorcyclists. Multiple
factors can contribute to this disproportionate impact, including environmental factors,
inattention of other motorists, driver error, and physical road characteristics. Many motorcycle
safety initiatives address the error and role of other drivers in motorcycle-involved collisions but
little attention is often given to the environmental and roadway factors. This lack of attention and
analysis reduces the ability of transportation agencies to obtain a complete common operating
picture for all factors impacting motorcycle safety – allowing for missed opportunities to
decrease the increased vulnerability of motorcyclists.
This project utilized Geographic Information Systems (GIS) as a tool to identify locations
on state-maintained roadways showing statistically significant clusters of motorcycle involved
collisions. The collision data for this report were retrieved from the Kentucky State Police
collision database; filters were used to extract motorcycle involved collisions for a ten-year
period from 2009 and 2018. A site suitability study was completed using the collision data and
road network layers to determine sites suitable to the introduction of a motorcycle lane in an
effort to increase motorcycle safety. While there are multiple strategies for reducing motorcycle
involved collisions, exclusive motorcycle lanes offer motorcyclists a safe location to ride without
interference from other motor vehicles in areas with high traffic flow. Spatial analysis was
utilized to complete a site suitability study to determine needed and viable locations for
motorcycle lanes using Livingston, Jefferson, Fayette, and Meade County within the
Commonwealth of Kentucky as a study area.

1

Chapter 1 Introduction
Motorcycle safety is a common transportation concern within the United States because of the
increased vulnerability for injury or fatality to the rider. Despite only accounting for 3% of the
vehicles on the road, motorcyclist fatalities comprise 14% of the total number of vehicle
fatalities (NHSTA 2019). Commonly used for commuting or recreational purposes, motorcycles
have many physical and structural differences when compared to other vehicles. These structural
differences, coupled with environmental and roadway factors, often allow for a greater amount
of injury to a motorcyclist if they are involved in a collision.
Multiple elements separate from vehicle design can increase the probability of a
motorcycle involved collision. Some impacting factors include the physical roadway
characteristics such as the roadway curve, pavement roughness, pavement type, curb presence,
and grade. Others include environmental variables including weather and time of day, other
motorist involvement, and negligence on the part of the motorcycle rider. The large number of
potential factors makes actionable analysis difficult to perform when attempting to predict the
probability or severity of a motorcycle involved collision. It also makes the creation of a
comprehensive geospatial view of safety considerations and possible hazards to motorcyclists
challenging and complex because of the vast amount of potential factors. Despite the difficulty
of creating a common operating picture of all impacting factors, motorcycle safety can still be
addressed and potentially improved by looking at existing crash data and then identifying the
factors present in those locations to determine what actions can be taken to improve safety.
Geographic Information Systems (GIS) were utilized as a tool within this thesis to
identify collision hot spots involving motorcycles occurring over a ten-year period, allowing for
potential locations needing infrastructure improvements to be identified. Once hot spots had been

2

identified according to certain parameters, additional analysis was performed to detect hot spots
overlaying a crash factor tied to the Average Annual Daily Traffic (AADT) amount. As will be
discussed in the literature review, the AADT of an area can indicate a higher presence of
motorcycle and other vehicle interactions, which can lead to higher probabilities of a motorcycle-
involved collision. This analysis was performed to identify locations that may be a suitable
location for the introduction of motorcycle lanes to the existing road infrastructure. Motorcycle
lanes can reduce the number of motorcycle-involved collisions, as the lane offers motorcyclists a
safe location to ride without interference from other motor vehicles. Spatial analysis was utilized
to complete a site suitability study to determine needed and viable locations for motorcycle lanes
in a specified area. The completion of this spatial analysis by this project can also help
transportation agencies identify and address roads that pose an increased risk to motorcycle
safety even if a motorcycle lane is not viable because of space constraints.
A foundation of knowledge for motorcycle risk on roadways was gained by reviewing
identified motorcycle crash factors, roadway engineering, and infrastructure improvement
studies, and existing mobile motorcycle applications. The findings from this review are included
within the Related Work chapter of this document. Case studies covering the introduction of
vehicle-specific travel lanes and motorcycle safety initiatives provided additional identification
of factors impacting motorcycle crash rates and viable countermeasures to reduce the risk of a
collision. The resulting information from a review of the above topics was utilized in guiding the
spatial analysis for this project within the designated study area. The study area for this project is
the Commonwealth of Kentucky due to data accessibility and researcher location.

3

1.1. Motivation
Motorcyclists are disproportionately injured or killed in crashes when being compared to
other motor vehicle occupants. As mentioned above, motorcyclist fatalities compromise 14% of
the total number of vehicle fatalities within the United States although they only account for 3%
of the vehicles on the road (NHSTA 2019). While this percentage fluctuates slightly from year to
year, the trend stays consistent: motorcyclists are more susceptible to injury or death resulting
from a collision. The motivation for this thesis stems from three inter-related components: the
natural vulnerability of motorcyclists, crash statistics and trends, and an overarching lack in
applied hazard analysis research on motorcycle safety.
1.1.1. Natural Vulnerability
Motorcycles have an inherent risk of greater injury to riders because of their design and
structure. They weigh less than other vehicles, contributing to crash severity and injury of the
rider in the event of a collision or crash (Rezapour 2019, 108). The smaller frame of a
motorcycle can make it difficult for other motorists to see – they are less visible and often hit as
a result. The majority of crashes involving motorcycles and other vehicles occur because the
other vehicles did not see the motorcycle, resulting in a reduced reaction time to successfully
redirect and avoid a collision (NHSTA 2019). Roads are also typically designed for automobile
use (Nabor 2016, 10) and certain aspects can increase the probability or impact of a collision on
a motorcyclist. Shoulder types (i.e. curb or flat), traffic control devices, curves, and pavement
conditions can increase the severity of a motorcycle involved collision despite being considered
safe or a necessary part of designing the roadway for general automobile use.
Motorcycles are less stable than other vehicles because of their smaller frame and design;
they offer a reduced amount of protection to the rider in the event of a collision (IIHS HLDI

4

2019). Part of the stability issue stems
from the two-wheel design –
motorcycles have a higher degree of
instability because of the two-wheel
design as opposed to the four-wheel
design of other motor vehicles. This
instability gives motorcyclists a higher
level of vulnerability while riding that
their counterparts in four wheeled
vehicles may not be susceptible to - the
impact of variations in road geometries
and surface conditions can increase the
probability of a motorcyclist losing
control and having a collision (Daniello
et al. 2010, 27). Figure 1 references the
common differences between motorcycles and other motor vehicles. This image was a part of a
motorcycle safety campaign from the Texas Department of Transportation and it helps exemplify
the structural differences which increase the vulnerability of the rider (Texas Department of
Transportation 2019). The natural vulnerability and reduced amount of protection to the
motorcycle rider resulting from the motorcycle design is part of the motivation for analyzing
motorcycle collisions within this thesis.
Figure 1. Motorcycle comparison

5

1.1.2. Statistics and Trends
Compounded with the inherent vulnerability of the motorcycle as a vehicle type because
of the structural design, motorcycles offer a reduced amount of physical protection to riders . The
smaller frame, lack of seatbelt and airbags, and lack of doors and a roof decrease the physical
protection to riders and their passengers in the event of a collision (Texas Department of
Transportation 2019). The reduced physical protection contributes to a higher injury and fatality
rate when compared to the injury and fatality rates of other vehicle occupants. The Highway
Loss Data Institute (HLDI) within the Insurance Institute for Highway Safety (IIHS) reported
5,172 motorcyclist fatalities in 2017 (HLDI IIHS 2018) and in 2016, they reported the number of
motorcyclists killed on roadways was 28 times higher than the number of deaths for other
vehicular traffic. The National Highway Traffic Safety Administration (NHSTA) elaborated on
the 2017 fatality total, stating motorcyclist deaths account for 14% of traffic deaths despite
motorcycles only comprising 3% of vehicles on the roadway. These statistics indicate a
disproportional relationship between expected vehicle motorist fatalities and actual motorcyclist
fatalities. The historically disproportionate impact on motorcyclists is part of the motivation for
evaluating motorcycle involved collisions within this thesis.
1.1.3. Existing Hazard Analysis
The existing hazard analysis for motorcycle collisions identifies a myriad of factors
which can impact motorcycle safety on roadways but the application of mitigation strategies to
reduce the impact of those factors is not as pronounced. Identifying factors impacting motorcycle
safety is crucial in reducing the risks and hazards motorcyclists face but action still needs to be
taken to increase the safety of these riders through mitigation strategies. The application of
mitigation strategies including physical initiatives and roadway alterations to address identified

6

factors impacting motorcycle safety is minimal within existing literature. It is necessary
transportation agencies are aware of the hazards and dangers their roadways pose to
motorcyclists but the next step of addressing those factors is critical to improve motorcycle
safety.
Aligned with the importance of transportation agencies being able to identify factors
having an impact on motorcycle safety, it is vital for motorcyclists on the roadways to be able to
identify potential hazards as they are riding. The concept of hazard perception addresses the
ability of a rider to perceive a threat and avoid impact while riding by altering their actions
(Cheng 2011). Motorcycle riders who had experienced collisions were shown to have a lower
level of risk perception, which may have contributed to their collision(s). Contrarily, motorcycle
riders who have not experienced a collision were shown to have a higher level of risk perception.
The concept of risk perception by motorcyclists lends support to the need to address potential
hazards and crash factors when they are identified – the impetus of applying mitigation
techniques to known collision factors to increase safety and reduce vulnerability.
Factors which can impact the probability of a motorcycle involved collision occurring
have been identified and will be further covered in the literature review section of this document.
The benefit of this project is offered in the application and use of those factors to determine
roadways that could benefit from motorcycle lanes and / or additional action by the
transportation cabinets. This project explores the identified factors impacting motorcyclist safety
and creates a geospatial product that displays hot spots for motorcycle involved in collisions
coupled with factors shown to have an impact on motorcycle safety. The research combined
existing data on motorcycle collisions, data on factors impacting motorcycle collisions, and
mitigation strategies in an effort to begin addressing the gap within the existing literature.

7

1.2. Study Area
Four counties within the Commonwealth of Kentucky were selected as viable study areas
for this project: Livingston County, Meade County, Fayette County and Jefferson County.
Visible in Figure 2, these counties were selected as a result of analyzing motorcycle collision
data between 2009 and 2018 collected from the Kentucky State Police’s (KSP) crash information
website (KSP, 2019). For the ten-year period, these counties experienced either a higher
cumulative collision amount, a higher probability of a collision resulting in an injury, or a higher
probability of a collision resulting in a fatality. The four counties were analyzed as separate study
areas to ensure the hot spots within that county were identified without interference or influence
from a different county’s collision records. This assisted in evaluating motorcycle safety and the
completion of this project’s objective. Although the study areas are specific to Kentucky, the
methods used for determining site suitability can be used for locations outside of Kentucky.
Figure 2. Commonwealth of Kentucky

8

1.2.1. The Commonwealth of Kentucky
The Commonwealth of Kentucky was selected as the location for research because of
data accessibility and researcher location. Located in the east central portion of the United States,
Kentucky has over 3.5 million registered vehicles including 94,675 registered motorcycles
(1.23%) (DataMart 2019). Comprised of 120 different counties split between 12 Highway
Districts, Kentucky is home to a variety of geographic features spanning the 27,500 + miles of
state-maintained roads. A portion of these state-maintained roads are named Scenic Byways
because of the “scenic, natural, cultural, historical, archaeological, and/or recreational” nature of
the roadway viewshed (DGI, 2019). The recreational nature of these routes can make them
preferred travel routes for recreational motorcycle riders and they, along with the remaining
state-maintained roadways, were evaluated within this project.
The average for motorcycle involved fatalities is lower in Kentucky as a whole when
compared to the 2017 national average of 14%, as referenced by a ten-year spread of fatality
counts shown in Figure 3. Despite a lower fatality average, motorcyclists are still
disproportionately represented in injury and fatality counts as shown in Figure 4. Between 2009
and 2018, motorcycle involved
collisions accounted for 1.18%
of all vehicular collisions in
Kentucky. The injuries resulting
from motorcycle involved
collisions, however, represented
a higher percentage of overall
injuries at 3.6% and an even
higher percentage of fatalities at 11.5%. The figures below help demonstrate the increased and
Figure 3. Motorcycle representation

9

disproportional vulnerability for individuals impacted by motorcycle involved collisions. The
study area focus was narrowed down from all of Kentucky to the specific counties having higher
percentages of overall collisions, injuries, or fatalities over the ten-year spread of collision data.
Once the study area was reduced to four specific counties, the factors shown to impact
motorcycle crash rates and existing crash data were evaluated in support of the objective of this
paper.
Figure 4. Kentucky crash fatalities
1.2.2. Livingston County, Kentucky
After reviewing the crash data, motorcycle-involved collisions were more likely to lead
to an injury within Livingston County – prompting the inclusion of the county as a study area.
Motorcycle collisions in Livingston County between 2009 to 2018 resulted in a higher
percentage of injuries when being compared to the total number of motorcycle involved
collisions on the county level. Kentucky State Police recorded 57 collisions involving
motorcycles, resulting in 60 injuries (Kentucky State Police 2019). There were a higher number

10

of injuries recorded than there were collisions, meaning multiple injuries resulted from collisions
during this time period. Comparatively, while Livingston County did not have the highest
percentage of injuries for Kentucky, the county did have the highest percentage of injuries
coupled with the highest injury count. Refer to Figure 5 for a chart of the counties with the
highest injury percentage stemming from motorcycle involved collisions.
Figure 5. Highest county injury percentage
Livingston County is located in western
Kentucky and is a part of KYTC District 1. The
county has a low population with 9,519 residents and
compromises 342.32 square miles (DataMart 2019).
There are 416.7 miles of state-maintained roadways
which were evaluated within this project. The
Kentucky Transportation Cabinet has 8,764 registered
vehicles within the county and 268 of those
registrations are motorcycles. Despite Livingston
County being one of Kentucky’s smaller counties,
benefit can be gained by using the county as a Figure 6. Map of Livingston County

11

study area. Livingston County has a higher percentage of injuries resulting from motorcycle
involved collisions and can benefit from an analysis of existing crashes and identified
motorcycle crash factors. Refer to Figure 6 for a map of Livingston County and the state-
maintained roadways within the county.
1.2.3. Meade County, Kentucky
Similar to Livingston County, Meade County was included as a study area because of the
higher percentage of motorcycle involved collisions resulting in a fatality; the collision data
revealed individuals had a higher change of dying from a motorcycle involved collision in
Meade County. Motorcycle collisions in Meade County between 2009 and 2018 resulted in a
higher percentage of fatalities when compared to the total motorcycle involved collisions
recorded for the county. There were 17 fatalities for the 108 motorcycle involved collisions
recorded over the ten-year period (Kentucky State Police 2019). The percentage of fatalities
resulting from a collision for this county was 16% and while this is not the highest fatality
percentage for all counties within Kentucky, it is the highest percentage coupled with the highest
Figure 7. Highest county fatality percentage

12

recorded fatality count. Refer to Figure 7 for a summary of the six counties with the highest
resulting fatality percentage.
Located in KYTC Highway District 4,
Meade County, Kentucky has 24,580 registered
vehicles with 910 of those being registered
motorcycles (DataMart 2019). With a population of
28,602, the county covers 324.43 square miles and
is located in south central Kentucky. The county
has over 500 miles of state-maintained roadway
that were evaluated to determine locations that
could be addressed to improve motorcycle safety
and reduce the resulting injuries and fatalities from
motorcycle involved collisions. Refer to Figure 8
for a map of Meade County displaying the state-
maintained roads within the county.
1.2.4. Jefferson County, Kentucky
Jefferson County experienced the highest cumulative amount of motorcycle-involved
collisions between 2009 and 2018 and was included as a study area. In this ten-year period,
3,160 collisions involving a motorcycle were recorded, resulting in 2,306 injuries and 142
fatalities (Kentucky State Police 2019). As shown in Figure 9, 10, and 11, Jefferson County
experienced the highest cumulative amount of injuries, fatalities, and collisions.
Figure 8. Meade County

13

Figure 9. Highest collision count by county
Figure 10. Highest injury count by county

Figure 11. Highest fatality count by county

14

Jefferson County has the
highest population of the four study
areas with 741,096 people although
the county covers an area similar to the
other study areas: 397.61 square miles
(DataMart 2019). Located in KYTC
District 5, Jefferson County has
1,952.2 miles of state-maintained
roadway that were evaluated within
this study for identified motorcycle
crash factors. There are 571,473
vehicles registered to the county with
12,161 of those vehicles being
motorcycles. Refer to Figure 12 for a
map of Jefferson County.
1.2.5. Fayette County, Kentucky
Fayette County was selected as a viable study area because the county had the second
highest cumulative count of motorcycle involved collisions (Figure 9), resulting injuries (Figure
10), and resulting fatalities (Figure 11). In this ten-year period, 3,160 collisions involving a
motorcycle were recorded, resulting in 1,271 injuries and 36 fatalities (Kentucky State Police
2019). It is important to note Fayette County was not originally included as a study area but was
selected after analysis had begun because of inconclusive results for portions of the analysis
Figure 12. Jefferson County

15

conducted on Livingston and Meade counties. This will be explained further in the Results
chapter of this document.
Fayette County, located in
central Kentucky, has a population of
295,803 people and is the smallest of
the study areas with an area of 285.49
square miles (DataMart 2019).
Located in KYTC District 7, Jefferson
County has 920.1 miles of state-
maintained roadway that were
evaluated within this project. There are
219,100 vehicles registered to the
county with 4,812 of those vehicles
being motorcycles. Refer to Figure 13
for a map of Fayette County

1.3. Thesis Organization
The following chapters within this document are designated for overarching components of work
completed for this project. The introduction and foundation have been laid in Chapter 1, where
the motivation and crux of the problem this project addressed was introduced: is it possible to
determine locations on state-maintained roads needing mitigation measures to increase
motorcycle safety? Chapter 2 looks at related work to motorcycle safety and the critical
components needed to be understood for a comprehensive, actionable analysis. The third chapter
Figure 13. Fayette County

16

of this document addresses the research methods, identifying and justifying the tools used for
analysis and outlining the procedures followed. Chapter 4 reveals the overall results per study
area, defining the issues encountered and limitations of the findings. The final chapter, Chapter
5, contains the conclusion, application, considerations, and next steps for this project.

17

Chapter 2 Related Work
While research exists relating to motorcycle safety, there is an identifiable gap in the application
of that research to efforts aimed at making roadways safer for motorcyclists. A plethora of
research on potential factors impacting a motorcycle’s traversement of roadways and safety
measures that can reduce motorcyclist vulnerability exists. There are multiple mobile
applications in production for Apple and Android devices a well geared toward increasing
motorcycle awareness and discovering / sharing safer riding routes. The research mentioned
above starts to create a picture of what motorcycle safety is and the various factors which impact
it. There is still a gap, however, in literature on combining the motorcycle mobile applications,
crash statistics, collision factors, and knowledge of mitigation measures for the improvement of
roadway conditions linked to motorcycle crashes. The existing literature has few examples
covering the application of safety measures to identified hazards.
2.1. Identified Factors
Identified factors in motorcycle collisions include roadway characteristics, other motorist
involvement, and negligence on the part of the motorcycle rider. The case studies identifying
motorcycle crash factors contain various study areas, some with inherit differences when
comparing the location to the project study area of Kentucky. These studies still provide valuable
insight into the factors affecting motorcycle safety. It is also important to note multiple studies
follow an approach of separating crash statistics into specific categories: motorcycle only
incidents, motorcycle and a single vehicle, and motorcycle and multiple vehicle collisions. This
delineation into data categories was adopted within this project as the factors impacting
collisions vary slightly based on the category of collision.

18

A prominent factor impacting motorcycle collisions is other drivers. In Victoria,
Australia, a study identified multiple factors as playing a role in motorcycle collisions with the
primary factor being other road users (Allen 2017, 157). Additional factors include rider age, the
traffic density of an area, speed of the rider and other vehicles, and road design issues. While it is
difficult to account for the human element of other drivers in a predictive, geospatial context,
analyzing the Average Annual Daily Traffic (AADT) can be useful. AADT reflects the average
traffic density on roadways – it is used to account for higher traveled roads where there is an
increased interaction between motorcyclists and other drivers.
It is important to note additional research suggests a variable use of AADT as a factor
impacting motorcycle crashes. There are multiple studies pointing to traffic density and
interactions with other drivers as a primary factor in motorcycle collisions (Sohadi 2000, 11) but
a contrary viewpoint needs to be considered (RideApart 2018): an inverse relationship to traffic
density and motorcycle crashes can be found in rural areas. The logic is that in lower traffic
density areas, motorcycles are not expected to be common and consequently are involved in
collisions because other vehicular traffic did not stop to look twice for them and they crossed the
path of the motorcycle. This clarification is important because it can affect the analysis
performed on high density and low density roadways when working with the traffic density or
AADT variable. Within the context of this project, the AADT for a roadway is analyzed to
determine the severity of the increased interaction between motorcyclists and other drivers.
A case study in Wyoming reviewed factors impacting vehicles in downgrade collisions
(steep downward slopes and hills / mountains) and identified lane width and speed as viable
factors (Rezapour 2019, 115). Despite the focus of this case study being on mountainous areas,
the results can still be used because it highlights factors which need to be monitored in areas with

19

variable grade. Additional motorcycle crash factors include roadway curve, presence of
driveways, two lane roads (Schneider 2012, 673), topography, presence of construction, road
geometries, railroad crossings, road condition, wildlife presence, and jurisdictional information
leading to population density (Ramirez 2016). Bridges along a roadway can also impact the
probability of motorcycle collisions as the bridge condition and approach affect the overall
rideability and ease of crossing the bridge (Murthy 1990). Uneven surfaces can enhance the
instability of motorcycles – a bridge approach in poor condition can knock a motorcyclist off
balance and contribute to a potential collision.
2.2. Motorcycle Specific Mobile Applications
Mobile applications are being included within the literature review as they provide
context for the applications used by the motorcycle community and the applications help identify
factors which can impact motorcycle safety. As motorcycle usage increases, it is important to be
aware of what goes into the selection of a route for riding and the considerations of a rider.
While it is difficult to find mobile motorcycle application in peer reviewed and scholarly
journals, web browser searches return a plethora of results for applications widely used within
the motorcycle community. Many applications can also be used outside of the motorcycle
community including gas availability applications, weather reports, and lodging applications.
These are often returned as “motorcycle” mobile applications because of the recreational nature
of motorcycle use. For the duration of this project, any reference to motorcycle mobile
applications will not refer to this last example of mobile applications.
Eat Sleep Ride is a mobile application combining functionality types for motorcycle
riders (BikeBandit 2018; Guido 2017; Gales 2017). It allows users to track, create, share, and
find existing motorcycle routes. The value from this application is the information collected and

20

stored with routes – the duration, level of difficulty, hazardous nature, etc. These route selection
factors begin to create an understanding of factors motorcyclists consider when determining a
safe route to ride.
A similar application, Rever, also facilitates route sharing (BikeBandit 2018; Guido
2017; Gales 2017) by allowing riders to record and share routes with various metrics. These
metrics include speed, distance, and elevation changes which were identified as factors
potentially impacting the probability of a motorcycle crash. A Road Segment Safety Rating
System application (Ramirez 2016) identified as a US Patent Application also assists
motorcyclists with finding safe routes. This application uses a rider-provided starting and ending
point and then calculates the safest route for the motorcyclist to travel. The application reviews a
set of road ratings and routes the motorcyclist on the route with the highest ratings. This
application can be further reviewed to determine how the roads are chosen and which metrics the
application uses.
2.3. Current Safety Initiatives
Current Safety Initiatives can help identify viable practices that increase or promote
motorcycle safety. Examining safety initiatives undertaken in various states and countries helps
identify the underlying causes impacting motorcycle involved collisions and it provides possible
mitigation practices for improving motorcycle safety. Hazards to motorcyclists can be segregated
into two categories: hazards stemming from the road itself – the infrastructure or environment
portion, and hazards stemming from behavior – and/or the rider or other motorists (Cheng 2011).
To assist with structuring a review on safety initiatives, the mitigation measures of safety
initiatives have been separated into similar categories for the behavior-based initiatives and
infrastructure-based initiatives.

21

2.3.1. Behavior-Based Initiatives
Behavior-based initiatives are meant to increase motorcycle safety by mandating or
encouraging changed behavior – they do not address changing the infrastructure of the road or
road assets but profess that through behavioral change, the level of safety for motorcyclists can
be increased. An example of a behavior-based initiative can be found in Australia. Motorcycle
crashes in Victoria, Australia, were evaluated to determine collision factors and multiple safety
improvement suggestions were provided (Allen 2017,165). These suggestions include white
helmets (a 33% reduction in crashes), reflective clothing (a 24% reduction in crashes), and
additional safety measures including re-evaluating posted limits and speed enforcement for riders
traveling over the speed limit. While these suggestions do not address the physical element of the
roads themselves, they do address factors which may lead to motorcycle involved collisions
including visibility and driving safety.
Road signage, media engagement and education, and inspection rides are additional
behavior-based initiatives pursued in various countries and states. In North Carolina, a segment
of roadway was evaluated in a Road Safety Audit (RSA) between 2012 and 2014 (Nabors 2016,
13). This segment of roadway had multiple recorded motorcycle involved collisions. As opposed
to changing the road infrastructure, signage specific to motorcyclists was added alerting them to
the road conditions (i.e. a sharp curve) which could impact them. In London, additional signage
was used for motorcyclists within work zones to alert riders to pavement changes, construction
debris such as loose gravel, and splash warnings (Nicol 2012, 8). Similar practices are used
within Norway, Belgium, and the United Kingdom where signs are placed for the benefit of
motorcyclists as well as other automobile drivers. In Norway, rides are organized with local
motorcyclists to inspect routes – these are called “road quality rides” and they are intended to
allow for reporting on pavement and road conditions, clear debris, and report any potential

22

concerns that may lead to a motorcycle involved collision. Engaging local groups as Norway did
is a successful example of a behavior-based initiative that allows for motorcycle safety and
awareness to increase without costly and potentially unneeded infrastructure changes.
Training classes are another layer of behavior-based initiatives that can lead to an
increase in motorcyclist safety and a potential reduction in motorcycle involved collisions. Basic
training courses are not required in all states within the United States and trends have shown that
most riders who do attend the basic courses do not continue their motorcycle safety education,
failing to take advanced courses (Nabors 2016, 10). When compared with other automobile
drivers, motorcyclists have a higher instance of having collisions with objects which are fixed
and not moving. Encouraging advanced training courses within an initiative could help reduce
this fixed object collision rate because of the maneuvers taught in the advanced courses. These
safety initiatives highlight factors which may impact the probability of a motorcycle involved
collision and they provide factors which can be monitored in areas containing either fewer or an
excess of motorcycle collisions.
2.3.2. Infrastructure-Based Initiatives
Infrastructure-based initiatives are meant to increase motorcycle safety by addressing the
infrastructure of the road and / or road assets and can be costly but needed and rewarding
endeavors. An example of an infrastructure-based initiative can be found in Norway (Nicol
2012) and North Carolina (Nabors 2016) where paved aprons are created where gravel or dirt
roads meet paved roads. An apron is a section of the roadway entrance where two roads meet. By
creating a paved entrance to the gravel or dirt road, there is a reduced amount of debris (dirt and
gravel) entering the traveled path of the paved road as vehicles turn and weather washes debris

23

across the road. If left uncleared or allowed to accumulate, debris on the roadways can create an
unstable riding environment and potential hazards for motorcyclists.
While signage was mentioned in the behavior-based section, modifying existing signage
is an infrastructure-based initiative aimed at increasing motorcycle safety for line of sight and
potential collisions. In Norway, road signs on multiple posts were consolidated to occupy one
post with multiple signs (Nicol 2016, 9). This can reduce visual obstructions which may distract
from or hide motorcyclists. Additionally, the construction of the sign posts was changed to a
lattice appearance to reduce the severity of injury in the event motorcyclists collided with
signage on the side of the road
In Malaysia (Radin et al. 2000, 11), traffic flow, also discussed within the factor section
of the literature review, is being used to justify the use of motorcycle lanes in a government
initiative to increase motorcycle safety. The majority of traffic fatalities (nearly 60%) within
Malaysia were motorcyclists and, despite the country having a higher percentage of motorcycle
riders, the impact to motorcyclists in collisions was disproportional (Law 2005, 3372).
Infrastructure changes were evaluated and the solution of motorcycle-exclusive lanes was
decided on. These lanes, approximately 3.81 meters (12.5 feet), were introduced on roadways
with over 60,000 AADT (Radin et al. 2000) in an effort to reduce collisions and the results were
successful. Opening a motorcycle lane was shown to reduce collisions by 39%, partly because it
reduced the interaction of vehicles and motorcycles traveling at different speeds. This example of
a current safety initiative lends support to traffic flow being a viable factor impacting motorcycle
collisions as well as motorcycle lanes being an appropriate response to an issue for motorcycle
safety. Within this project, motorcycle exclusive lanes were the infrastructure-based initiative
being reviewed for site suitability. This is further discussed within the Methods chapter.

24

Chapter 3 Data and Methods
The objective of this project was to determine suitable and viable locations for motorcycle lanes
on state-maintained roadways in Kentucky. This was accomplished by using GIS as a tool to
evaluate motorcycle collision locations from 2009 to 2018. After reviewing the data, several
study areas were selected within Kentucky. Several methodologies were considered and
evaluated to accomplish this objective. The first route evaluated involved using Esri’s Network
Analyst to assign a route a severity rating based on the presence of existing hazards from the
road infrastructure. This route was not pursued after the initial evaluation because the tool did
not allow for the identification of locations needing infrastructure improvement – it was more
suitable for determining routes with a higher or lower rating score based on multiple factors. The
second route evaluated for completing this project’s objectives involved using a different tool
(the Optimized Hot Spot Analysis Tool) and it is outlined below.
GIS was used to complete a technical hot spot analysis for collisions according to various
parameters. Identified hot spots were then evaluated to determine if they were located on state-
maintained roads and segments with an Average Annual Daily Traffic (AADT) value of over
60,000 were then isolated. The resulting roadway segments were evaluated for lane-widening
potential to determine which locations could support the infrastructure addition of a motorcycle
lane. The use of motorcycle lanes in collision-prone locations showing a higher AADT could act
as a viable motorcycle safety approach to reduce the number of motorcycle-involved collisions.
This reduction in motorcycle-involved collisions can also potentially reduce the number of
collision related injuries and fatalities.

25

3.1. Data Description
Multiple datasets were evaluated in determining suitable locations for motorcycle lanes.
The data being used in this project was grouped into three, theme-based categories: base or
jurisdictional information, crash factors, and recorded crash locations. Throughout the project,
different types of analysis were used for the three categories. The sections below identify the
data within the categories and the analysis performed in determining suitable locations for
motorcycle lanes.
3.1.1. Jurisdiction and Base Layers
Various jurisdictional boundaries and base data layers were used within this project for
summation information and visual display. The boundaries for the Commonwealth of Kentucky
and counties within Kentucky were collected as shapefiles from the Department of Geographic
Information’s (DGI) geoportal (2019) which acts as a platform for state and local agencies to
publish data layers. These layers are open to the public and maintained/updated by DGI. They
are both polygon layers and can be added to a map document, giving the ability to access
summation information, clip other layers, and allow for a visual representation of the study area.
Road centerlines were used as a data layer within this project for visual display, as a base
layer, and for road segment identification. Two separate layers were used: one to represent the
state-maintained roadways, and one to represent local-maintained roadways. Both layers were
obtained as shapefiles through the DGI geoportal. The Kentucky Transportation Cabinet (KYTC)
utilizes the geoportal to publish various data layers for public consumption. Table 1 displays
additional information on the jurisdiction and base layers which were used within this project.

26

Dataset Source Scale Precision Accuracy Purpose
Kentucky
Boundary
DGI NAD 1983
StatePlane
Kentucky FIPS
1600 Feet
Vector Polygon 95% Confidence
the accuracy is
between 0.5 and
2.0 Meters
Jurisdictional
boundary for visual
identification and
data management
County
Boundary
DGI NAD 1983
StatePlane
Kentucky FIPS
1600 Feet
Vector Polygon 95% Confidence
the accuracy is
between 0.5 and
2.0 Meters
Jurisdictional
boundary for visual
identification and
data management
Road
Centerlines
– State and
Local
KYTC
through
DGI
NAD 1983
StatePlane
Kentucky FIPS
1600 Feet
KYTC maintains data
collection standards for road
centerlines and associated
attributes to help increase
the precision of their data.
Vector line layer.
95% Confidence
the accuracy is
between 0.5 and
2.0 Meters
Used in analysis for
locational
information and
road network
references
3.1.2. Collision Factor Data Layer
While several variables were identified through the literature review as motorcycle
collision factors, only one variable was evaluated within this study as it has a direct connection
to a possible mitigation measure. Traffic Flow (TF) was evaluated because of the connection
between higher AADT values and the success of the introduction of a motorcycle lane. This
layer was made accessible to the public as shapefiles through the Kentucky Transportation
Cabinet’s (KYTC) Datamart website. The Traffic Flow layer being used within this project
contains base road attribution (road name, mile points, unique identifiers) and attributes specific
to traffic flow: Average Annual Daily Traffic (AADT), count (LASTCNT), and the year that
count was conducted in (LASTCNTYR).
The Rating Evaluation Section (EV) was used as a factor to determine viable route
segments that can support the addition of a motorcycle lane to the existing infrastructure. This
dataset, also provided as a shapefile to the public by KYTC, contains roadway attributes
Table 1. Jurisdiction and base layers

27

including the widening potential of roadways (WIDENFEAS) and potential widening obstacles
(WIDENOBST). Refer to Table 2 for additional information on the crash factor layers to be used
within this project.
Dataset Source Scale Precision Accuracy Purpose
Average
Annual Daily
Traffic
(AADT) (TF)
KYTC
through
DataMart
NAD 1983
StatePlane
Kentucky FIPS
1600 Feet
KYTC maintains data
collection standards for road
centerlines and associated
attributes to help increase the
precision of their data. Vector
line layer.
95% Confidence
the accuracy is
between 0.5 and
2.0 Meters
Identified crash factor –
used for analysis
Rating
Evaluation
Section (EV)
KYTC
through
Datamart
NAD 1983
StatePlane
Kentucky FIPS
1600 Feet
KYTC maintains data
collection standards for road
centerlines and associated
attributes to help increase the
precision of their data. Vector
line layer.
95% Confidence
the accuracy is
between 0.5 and
2.0 Meters
This dataset contains
attributes for widening
potential which will help
determine areas which
can support the addition
of a motorcycle lane
3.1.3. Recorded Crash Layer
Motorcycle crash locations spanning ten consecutive years within Kentucky were
evaluated as a variable in order to determine hot spots of collision activity. Kentucky State Police
(2019) (KSP) provides collision information for all vehicle types through a Kentucky Collision
Analysis website. The website is designed to let the public create a query and access resulting
crash records. For this project, crash records were queried using motorcycles as a vehicle type
and crash dates between January 1, 2009 to December 31, 2018. A ten-year period was selected
Dataset Source Scale Precision Accuracy Purpose
Crash
Locations
KSP Layer was
created using
WGS 84 and
then projected
to NAD 1983
StatePlane
Kentucky FIPS
1600 Feet
Records accessed
through excel
sheet download.
A vector point
layer was created
Unknown –
records accessed
through excel
sheet download
Core analysis data
layer – Contains
location of
collision and
collision attributes
Table 2. Collision factor layers

Table 3. Recorded collision data

28

to account for fluctuations between years due to factors such as weather. This information was
returned in an excel table format and was then imported into ArcPro as a table. The resulting
table was used to make a point layer with the latitude and longitude attribution inherent in the
excel download. Additional attribution included with the crash data is the county, roadway, mile
point, collision date and time, number of vehicles involved, number of persons injured, number
of persons killed, roadway condition, weather, collision manner, roadway characteristics, light
conditions, and factors involved. Supplementary information on the recorded crash layer can be
viewed in Table 3 below.
3.2. Research Design
To support the project objective, the data above was analyzed and compared using the
Environmental Systems Research Institute’s (Esri) ArcGIS Pro program. Figure 14, referenced
below, is a graphical representation of the overall workflow used in completing this project’s
objective. There were four overarching phases for this project: (1) configure layers and map
document, (2) analyze crash location layer using the Optimized Hot spot Analysis tool, (3)
identify suitable roadway segments based on state maintained roads, AADT, and widening
feasibility, and (4) summarize the viable locations which can sustain a motorcycle lane.

29

3.2.1. Configure Layers and Map Document
The first step in this project was to gather data identified through the literature review as
relevant to motorcycle crash factors, gather layers related to jurisdictional and base
transportation, and gather motorcycle collision data for the specified ten-year period. These
layers were then added to an ArcGIS Pro document and grouped based on their role in the
research as either jurisdictional / base information, collision factors, or collision data. ArcGIS
Pro was selected as the mapping program to use for various reasons. This project required
multiple map frames for each of the study areas, multiple layouts for exporting the map frames,
the ability to run processes efficiently, and the infrastructure to retain the history of all processes
or tools run. ArcGIS Pro, as an operating platform, was able to satisfy all the requirements above
while also having the benefit of an intuitive user interface and the ability to access and use Esri
4. Summarize Viable Motorcycle Lane Locations
Utilize Summary Tools within ArcGIS Pro
3. Identify Suitable Roadway Segments
Select Roadway Segments based on
Hotspot Analysis
Refine selection based on AADT
values exceeding 60,000
Refine selection based on widening
feasibility
2. Utilize Optimized Hotspot Analysis Tool
Vehicle Focus All Collisions Casuality Focus
1. Configure Layers and Map Document
Add and group jurisdictional and base
roadway information layers
Add and group crash factor layers
Add and configure motorcycle collision
layer
Figure 14. Methods outline

30

basemaps. ArcGIS Pro also stores its files in the form of projects, helping organize the products
used and created while completing the technical work. Additional mapping programs such as
QGIS or ArcMap satisfied some of the requirements listed above but not all of them; it is
because of this ArcGIS Pro was selected for use in satisfying the objectives of this project.
Once the shapefiles and table were imported to ArcGIS Pro, the shapefiles were
symbolized based on their attribute (i.e. route type, AADT level) to facilitate the visual
interpretation of the layers. The collisions records, imported as an excel table, were turned into a
point layer based on the X (Longitude) and Y (Latitude) attributes via the XY Table to Point
tool. Upon layer creation, the projection was set to WGS 1984 for this layer and then projected to
match the projection for the rest of the layers as NAD 1983 StatePlane Kentucky FIPS 1600
Feet.
This collision layer was then copied six times into the same map frame, allowing for
distinct queries to be applied to each copy of the master collision layer for the state. Refer to
Appendix A for a complete listing of all queries used per study area. These queried layers were
categorized into three groups: a master layer with all records, vehicle status layers, and injury
status layers. The vehicle status group contained three queried layers: motorcycle only collisions,
motorcycle and one vehicle collisions, and motorcycle and multiple vehicle collisions. The injury
status group contained three layers as well: no casualty collisions, collisions involving one or
more fatalities, and collisions involving one or more injuries (but no fatalities). Refer to Figure

31

15 for an example of the resulting table of content with all layers in their respective groups and
the selected symbology.
The decision was made to use different subsets of data for running the Optimized Hot
spot Analysis tool for two main reasons. First, the information and practices uncovered in the
literature review suggested different levels of factor involvement and severity can be determined
on the manner of the collision. For instance, higher levels of injury are found to result from a
road design issue (Allen 2017, 165), suggesting the road design in hot spots found for collisions
involving injuries or fatalities may have played a factor in the collision. The identification of hot
spots relating to injury severity can alert transportation agencies of a potential need to evaluate
the road design in that area. The vehicle involvement in a motorcycle collision can also assist in
Figure 15. Layers and symbology

32

determining the best mitigation strategies for reducing motorcycle involved collisions in an area
(Schneider 2012, 669). Different factors can contribute to the manner of motorcycle only,
motorcycle and one vehicle, and motorcycle and multiple vehicle collisions – completing the hot
spot analysis on these subsets of the collision records increases the potential of the results to be
utilized in an efficient and appropriate manner.
Creating subsets of the collision records also allowed for identifying hot spots which may
not have been identified as statistically significant otherwise. Using subsets allowed the analysis
to identify hot spots relating to collision severity and vehicle involvement that would not have
been visible in a hot spot analysis of all motorcycle involved collisions. For instance, hot spots
were identified for fatality involved collisions differing from hot spots for causality free (no
fatality or injury) collisions – this would not have been possible if the collision records were
evaluated as a complete set with all records. The potential implications of this varied hot spot
analysis are further discussed in the Conclusion section of this document.
Once all layers were queried appropriately, symbolized, and grouped, map frames were
made specific to the different study areas (Jefferson County, Fayette County, Meade County, and
Livingston County) and the base data layers and groups were copied to the new map frames. The
queries applied on the base data frame (specific to the entire state of Kentucky) were then
modified for the appropriate county, allowing for a focused data view specific to that study area.
Layouts (formatted map views with components including titles, legends, scales, etc…) were
then created for every map frame to facilitate the sharing and preserving of results from the
analysis and technical work.

33

3.2.2. Utilize Optimized Hot Spot Analysis Tool
The second overarching category of work within this research methodology was to
analyze the collision subsets. The Optimized Hot Spot Analysis (OHSA) tool was chosen to be
used on each of the collision subsets to support one of the projects objectives of identifying areas
that could benefit from the introduction of a motorcycle lane by identifying areas with
statistically significant clusters of motorcycle involved collisions. The OHSA tool uses the Getis-
Ord Gi statistic to analyze records within the input layer (i.e. the motorcycle collisions) to
determine if there are statistically significant clusters (Esri 2019). Statistically significant clusters
indicate the possibility of collision occurrences which are not random and may be the result of
factors present within that location.
The Optimized Hotspot Analysis tool or the Getis-Ord GI Statistic has been used in
multiple case studies to determine areas of statistical significance for different phenomena. For
instance, the tool was used to determine hot and cold spots for power outages within certain
cities in California (Sultan et al. 2016, 229). Further analysis was then completed by researchers
on the identified areas of statistical significance to evaluate the age of infrastructure as a factor in
outages. The Getis-Ord GI statistic was also used in identifying traffic accident hot spots for
Brunei Darussalam, a country in South East Asia (Zahranel-Said et al. 2019, 1). The Getis-Ord
GI statistic allows researchers to evaluate the occurrence or frequency of widespread
phenomena, such as power outages, crime or accidents. The tool is used strictly for frequency,
however, and cannot take into account severity (unless the data is already queried or formatted to
account for severity). Within this project, areas of statistical significance were identified using
the Getis-Ord GI Statistic and then those areas were further evaluated for the presence of factors
relating to motorcycle-involved collisions.

34

The tool creates an output layer with several attributes representing the level of statistical
significance for each cell of analysis: the Gi_Bin, z-score and p-value. The Gi_Bin is the field
representing the percentage of statistical significance (3 = 99% Confidence, 2 = 95%
Confidence, 1 = 90% Confidence, and 0 = no statistical significance) (Esri 2019). The z-score
value for the record indicates the amount standard deviations the cell is away from being
considered random. A value of 1.65 indicates the cell has reached the 90% confidence level, a
value of 1.96 indicates a 95% confidence level, and a value of 2.58 or greater indicates at least a
99% confidence level of the cell not being spatially random (Law and Collins 2016, 323). While
the z-score represents the standard deviations for a cell, the p-value records the probability of the
records within a cell being random. A value of 0.1 indicates the cell has reached the 90%
confidence level, a value of 0.05 indicates a 95% confidence level, and a value 0.01or smaller
indicates at least a 99% confidence level of the cell not being spatially random – the closer the
value is to one, the higher the probability the records are spatially random (Law and Collins
2016, 324).
The Optimized Hot Spot Analysis tool requires several parameters be set before it can run
on the input dataset and it allows for certain parameters to be modified from the default. For this
project, the parameters were set consistently for all study areas and motorcycle collision subsets.
Prior to running the tool, a file geodatabase was created within the ArcGIS Pro Project folder to
house all resulting layers from the analysis. The tool was then populated by defining the input
layer, specifying the output layer name and location within the file geodatabase, setting the
aggregation method to the “Count incidents within fishnet grid” selection, selecting the queried
county layer specific to the study area, and overriding the cell size following a 5000 meter, 500
meter, 250 meter, and 100 meter scale for each run of the tool (the tool was run multiple times on

35

each layer for the various cell sizes). Figure 16
displays the tool parameters for the Fayette
County master collision layer at a 250 meter
cell size. It is important to note that while the
OHSA tool was utilized for every motorcycle
collision subset layer, all cell sizes were not
attempted for every layer – this will be
explained further within the Results section of
this document. Refer to Appendix B for a table
containing the layers and which cell sizes were
attempted through the OHSA tool.

3.2.3. Identify Suitable Roadway Segments
Once the OHSA tool was completed for the motorcycle collision subsets for every study
area, roadway segments suitable to the introduction of a motorcycle lane with statistically
significant clusters of motorcycle involved collisions were identified. This third overarching
category of work within the research methodology was completed by using various tools and
queries. The purpose was to identify locations with statistically significant collision clustering on
roadways adhering to the following parameters:
(1) the route must be a state-maintained road
Figure 16. OHSA example

36

(2) Average Annual Daily Traffic (AADT) value of at least 60,000
(3) existing infrastructure capable of being widened to support a motorcycle lane
The original intention was to evaluate each subset of motorcycle collision hot spot results
by using the union tool to combine the hot spot layers into a single hot spot record per subset of
vehicle involvement and injury status for the specific study area. The hot spot layers for
motorcycle only, motorcycle with one vehicle, and motorcycle with multiple vehicles would
have been combined to a single hot spot layer on the vehicle involvement. Similarly, the hot spot
layers for no causality, fatality involved, and injuries without fatalities would have been used
within the Union tool to create a single hot spot area representing the injury status of collisions.
A uniform cell size could not be used on all layers for the OHSA, however, and this planned
route of analysis could not be completed. The master collision layer for each study area was used
for identifying suitable roadway segments and that process is outlined below.
The queries were set first for each study area to ensure the parameters above were met
and factored in for the analysis to determine site suitability. Queries were set on the resulting
OHSA layers to filter for cells with 90% confidence of statistical significance using the Gi_Bin
attribute (values not equal to 0). Queries were then set on the Traffic Flow (TF) layer for a
LASTCNT value equal to or greater than 60,000, and the Rating Evaluation Section (EV) layer
for a D_WIDENFEA value not equal to “No widening is feasible.” Once these queries were set,
Esri geoprocessing tools were used to identify roads with the above attributes.
The Clip Layer tool was used first to determine if any of the identified hot spots were
located on a state-maintained roadway. The input layer was the state roads layer queried for the
specific study area. The clip layer used for the tool for each study area was a viable and complete
OHSA layer for all collisions within the study area – this will be discussed in greater detail

37

within the Results section. The output for the tool was saved within the file geodatabase for the
project. Refer to Figure 17 for an example of this process in Jefferson County. The output from
this tool was then added to a new group within the map frame titled Results and the symbology
was adjusted to be a green line of width 4 pts. This specification was decided on because it made
the output easily identifiable when viewed with other layers on the map.
It is important to note the Traffic Flow layer and the Rating Evaluation Section layer used
within this project were collected by KYTC on state-maintained roads. These two layers were
intentionally being used because they are factors in determining the site suitability of a potential
motorcycle lane. Consequently, the roads being evaluated within the scope of this project for
their capacity to house an additional lane exclusively for motorcycles are only state maintained
roads. Motorcycle collision records were included within the project regardless of occurring on a
state or locally maintained road to ensure the hot spots identified through technical analysis were
statistically significant and indicative of areas sincerely needing infrastructure improvements for
improving motorcycle safety.
After a successful completion for the Clip Layer tool, the Clip tool was used to determine
if any state-maintained roads within the OHSA layer also had an AADT value equal to or
exceeding 60,000. The input for this tool was the output from the Clip Layer tool and the clip
features were the Traffic Flow layer specific to the study area (refer to Figure 18 for an example
of this tool being applied to Jefferson County). Upon a successful completion of this tool, the
layer was added to the Results layer group and the symbology was updated to a yellow line of
width 2.5 pts. This allowed for stacking of the tool outputs, allowing the user to view the
difference in viable routes based on AADT. The final tool used in the analysis was another clip
tool. The output from the last tool was used as the input for this tool and the final clipping

38

features were the Rating Evaluation Section (refer to Figure 19 for an example of this tool being
applied to Jefferson County). This final step allowed for the identification of state-maintained
roads within an identified hot spot with an AADT value exceeding 60,000 and the potential to
have an additional lane exclusive to motorcycles added.
Figure 17. Clip layer tool Figure 18. Clip for AADT
Figure 19. Clip for widening

39

3.2.4. Summarize Viable Motorcycle Lane Locations
The fourth and final overarching category of work
within this research methodology was to create a summary
table from the final output of viable locations, allowing
for the information to be shared with key stakeholders as a
summary report of findings for the study area. The Esri
Summary Statistics tool was used to accomplish this task,
allowing for the creation of a table specifying the different
routes determined through analysis to be a viable location
as well as the count for how many times that route was
identified and the total length of roadway directly within
the hot spot. The final layer was used as the input for the
tool and attributes within the layer were selected with the
summary information needed. This summation table was
added to the final layouts for the study area. Refer to Figure 20 for an example of the summary
statistic table created for Jefferson County.

Figure 20. Summary statistics

40

Chapter 4 Results
The methodology outlined in Chapter 3 was completed for each of the study areas although the
entire methodology could not be completed on Livingston and Meade counties because
motorcycle lane parameters could not be met by the county infrastructure and need. The
motorcycle-involved collision hot spots, state-maintained roadways, Average Annual Daily
Traffic (AADT), and lane widening potential were analyzed for each study area and if a county
did not have locations which could satisfy the parameters outlined in Chapter 3, the analysis was
ended inconclusively. Appendix C contains a table referencing the results for the complete
analysis on each study area. Locations with the need and infrastructure for the addition of an
exclusive motorcycle lane were only able to be identified within Fayette and Jefferson County.
4.1. Livingston County Results
Throughout the technical work it was discovered Livingston County did not meet all parameters
for determining the site suitability of a motorcycle lane. The county does not have locations
where the AADT exceeded 60,000 within identified hot spots and consequently was not
evaluated on the infrastructure element of widening feasibility. The need for a motorcycle lane
could not be justified.
The Optimized Hot Spot Analysis tool was successfully completed on the master
collision layer for the county as well as the queried layers representing motorcycle only involved
collisions and injury only involved collisions. The cell sizes used, results, and any errors
received can be seen within Table 4. Multiple hot spots were found within the county for the
different subsets and the master layer for motorcycle involved collisions (refer to Appendix D for
maps of the varying hot spots). State-maintained roads within the hot spot areas were identified
(refer to Figure 21) using the Clip Layer tool as the methodology outlined but the following step

41

of using the Clip tool to determine which of those routes had an AADT value equal to or greater
than 60,000 yielded no results and the analysis was concluded for Livingston County (Table 4).

Study Area Focus Area # of Records100m 250m 500m 5000m **Z m OHSA Results Final Results
Multiple Vehicles 1 NC NC NC NC NC *Failed
Single Vehicle 17 NC NC NC NC NC *Failed
Motorcycle Only 39 NC NC CV NC CI Complete
No Casualty 10 NC NC NC NC NC *Failed
Fatality Greater than or equal to 1 1 NC NC NC NC NC *Failed
Injury greater than or equal to 1 46 NC NC CV NC CI Complete
All County Collisions 57 NC NC CV NC CI Complete
CV
Completed, Viable Result
NC
Not Completed
CI
Completed, Inconclusive
*Esri generated error code (001570) signifying the layer did not have
the minimum requirement of 30 records and the OHSA Tool failed
** Tool was completed using automatically generated cell size
No viable locations identified -
county does not contain state-
maintained roads meeting or
exceeding 60,000 AADT
intersecting the identified hot
spots.
Livingston
County, KY
Table 4. Livingston County results
Figure 21. Livingston County final analysis

42

4.2. Meade County Results
Meade County did not meet the parameters for determining the site suitability of a motorcycle
lane as the county does not have any state-maintained routes with an AADT at or exceeding
60,000. Within this project, motorcycle lanes are viewed as a potential mitigation measure to
increase motorcyclist safety for areas with traffic volumes high enough to warrant them. Similar
to Livingston County, Meade County was evaluated for hot spots, their location on state-
maintained roads, and then the AADT values were checked. Once it was determined, the AADT
parameter could not be met, the analysis stopped, and the infrastructure element of widening
feasibility was not evaluated. The need for a motorcycle lane could not be justified despite hot
spots being successfully identified.
The Optimized Hot Spot Analysis tool successfully completed for the Meade County
master collision layer and the queried layers for a motorcycle and single vehicle collision,
motorcycle-only involved collisions, and injury only involved collisions. Similar to Livingston
County, varying cell sizes were used to gain successful completions of the OHSA tool. The cell
sizes used, results and any errors received can be seen within Table 5 (refer to Appendix E for
maps of the varying hot spots). The state-maintained roads within the county were successfully
identified (refer to Figure 22) using the Clip Layer tool but the state-maintained roads within
Study Area Focus Area # of Records100m 250m 500m 5000m **Z m OHSA Results Final Results
Multiple Vehicles 3 NC NC NC NC NC *Failed
Single Vehicle 44 NC CV CI NC NC Complete
Motorcycle Only 61 NC NC CV NC NC Complete
No Casualty 25 NC NC NC NC NC *Failed
Fatality Greater than or equal to 1 16 NC NC NC NC NC *Failed
Injury greater than or equal to 1 71 NC NC CV NC NC Complete
All County Collisions 108 CI CI CV NC NC Complete
CV Completed, Viable Result
NC Not Completed
CI Completed, Inconclusive
*Esri generated error code (001570) signifying the layer did not have
the minimum requirement of 30 records and the OHSA Tool failed
** Tool was completed using automatically generated cell size
No viable locations identified -
county does not contain state-
maintained roads meeting or
exceeding 60,000 AADT.
Meade
County, KY
Table 5. Meade County results

43

Meade County all have AADT values lower than the 60,000 needed for analysis and the analysis
was inconclusive for determining the site suitability of a motorcycle lane.

4.3. Jefferson County Results
Jefferson County met all parameters for this study, and multiple locations were identified as
viable sites for the introduction of a motorcycle lane. In following the methodology outlined in
Chapter 3, hot spots were identified within the county for the master collision layer and suitable
roadway segments were identified for locations meeting all parameters (state-maintained, AADT
exceeding 60,000, with widening potential). Multiple sections of roadway met these parameters
Figure 22. Meade County final analysis

44

throughout Jefferson County and a summation table was created to capture the roadway name,
number of sections, and overall length of the sections in feet (refer to Figure 23).
Unlike Meade and Livingston counties, the Optimized Hotspot Analysis tool was able to
successfully complete for the master collision layer and every subset of records. The resulting
maps can be viewed within Appendix F, and Table 6 contains the results and cell size for each of
the layers. The hot spot layer produced from the master collisions dataset for Jefferson County
Figure 23. Jefferson County final analysis

45

was used for identifying suitable roadway segments. The Clip Layer tool was used to select the
state-maintained roadways within the statistically significant areas and the AADT value of
60,000 was used to further clip the roadway segments. The widening feasibility was evaluated
and 70 roadway segments were identified, predominantly on I-264, I-64, and I-65. All aspects of
analysis outlined within the methods chapter were successfully used for the Jefferson County
study area and areas with need and the infrastructure to support a motorcycle lane were
identified.

4.4. Fayette County Results
The Fayette County study area also met all parameters, and the outlined methodology was
utilized to successfully identify sites with the need and infrastructure to support the addition of a
motorcycle lane. The master collision layer for Fayette County was used in identifying suitable
roadway segments meeting the parameters as state-maintained, having an AADT value at or
exceeding 60,000, and having the potential to be widened. Similar to Jefferson County, Fayette
County had multiple sections of roadway meet the parameters above. These segments were
summarized by road name, the number of occurrences, and the total length of the segments in
feet (refer to Figure 24).
Table 6. Jefferson County results
Study Area Focus Area # of Records100m 250m 500m 5000m **Z m OHSA Results Final Results
Multiple Vehicles 157 CI CV CV NC CI Complete
Single Vehicle 2099 CI CV CI NC CI Complete
Motorcycle Only 904 CV CI CI NC CI Complete
No Casualty 1109 NC CV NC NC CI Complete
Fatality Greater than or equal to 1 141 NC CV CV NC CI Complete
Injury greater than or equal to 1 1951 CV CI NC NC CI Complete
All County Collisions 3160 CV CI CI NC CI Complete
CV
Completed, Viable Result
NC
Not Completed
CI
Completed, Inconclusive
*Esri generated error code (001570) signifying the layer did not have
the minimum requirement of 30 records and the OHSA Tool failed
** Tool was completed using automatically generated cell size
Multiple locations with need
and infrastructure capable of
supporting a motorcycle lane
identified.
Jefferson
County, KY

46

The Optimized Hotspot Analysis tool was able to successfully complete the master
collision and all collision subsets for the Fayette County study area. The resulting maps can be
viewed within Appendix G and Table 7 contains the OHSA metrics and the final result. Parallel
to the other study areas, differences in cell size for the OHSA tool led to the master collision
OHSA output being used. The remainder of the analysis steps were completed and 29 roadway
Figure 24. Fayette County final analysis

47

segments were identified throughout Fayette County. Interstate 75 had the highest occurrence
within the results and the highest cumulative length of viable segments. The potential need (areas
of statistical significance and AADT at or exceeding 60,000) and infrastructure to support the
implementation of motorcycle lanes were validated with the analysis for Fayette County.

It is important to note Fayette County was not included within the original scope of this
project. The county was added as a study area once inconclusive results were found in both
Livingston and Meade County. The initial analysis showed Fayette County as a potential study
area that would lend to complete results because of the volume of recorded collisions within the
county – Fayette County had the second-highest count for cumulative collisions, fatalities, and
injuries. In an effort to apply the complete methodology to a second study area, the county was
added to the project.

Table 7. Fayette County results
Study Area Focus Area # of Records100m 250m 500m 5000m **Z m OHSA Results Final Results
Multiple Vehicles 50 NC CI CV NC CI Complete
Single Vehicle 837 NC CV NC NC CI Complete
Motorcycle Only 384 NC CV NC NC CI Complete
No Casualty 506 NC CV CI NC CI Complete
Fatality Greater than or equal to 1 36 NC CI CI NC CI Complete
Injury greater than or equal to 1 735 NC CV NC NC CI Complete
All County Collisions 1271 CV CI CI NC CI Complete
CV
Completed, Viable Result
NC
Not Completed
CI
Completed, Inconclusive
*Esri generated error code (001570) signifying the layer did not have
the minimum requirement of 30 records and the OHSA Tool failed
** Tool was completed using automatically generated cell size
Multiple locations with need
and infrastructure capable of
supporting a motorcycle lane
identified.
Fayette
County, KY

48

Chapter 5 Conclusion
The objective of this project was to create a reproducible methodology and successfully use GIS
as a tool in determining the site suitability of motorcycle lanes within Kentucky to help increase
motorcycle safety and save lives. This objective was met, and viable motorcycle lane locations
were identified with a methodology that used multiple tools within ArcGIS Pro for hot spot
analysis, querying, clipping, and summarizing. The study areas of Fayette County and Jefferson
County, Kentucky, produced viable locations while the study areas of Meade County and
Livingston County produced inconclusive results relating to traffic flow volumes. The gap
identified within the literature review was also addressed within this process. A connection was
made between identified factors and viable mitigation measures which could act as a solution to
addressing the identified hazard. The methodology followed within this project provides a
framework for further evaluation of factors identified as impacting motorcycle involved
collisions and viable mitigation measures which could reduce the factor’s impact. The
methodology allows future work to continue closing the gap identified within the literature
review of this project.
Although the Livingston and Fayette results were inconclusive, they were not unforeseen.
Motorcycle lanes, as described within the related work chapter of this document, are viable
solutions in areas where there are high volumes of motorcyclists and high volumes of traffic.
Meade and Livingston County are not as populated as Jefferson and Fayette counties; they also
have lower average AADT values, and the cost of modifying the existing infrastructure to create
a motorcycle lane may not be justified because the population is not there to use it. Alternative
solutions, including widening the lane or having targeted motorcycle safety campaigns, may be
more feasible if the collisions are related to traffic flow or inattention of other drivers. Increasing

49

the lane width has been shown to decrease the severe injury risk by over 27 % (Rezapour 2019,
115) because motorcyclists then have a larger amount of room for maneuvers attempted in
avoiding imminent collisions. Reviewing the lane-splitting legislation, as discussed in the related
work chapter, in an area with higher collision rates may also be a viable alternative to changing
the existing infrastructure of the roadway for the creation of a motorcycle lane.
5.1. Project Limitations
When discussing the limitations of the work completed for this project, it is important to
differentiate between what was expected to be completed and what was not able to be completed.
Throughout the analysis and technical work, several expectations or intentions were set: create a
replicable process, utilize the collision subsets for a focused analysis, and evaluate multiple
identified collision factors. All of these expectations were met in varying capacities, although
they were not all met completely. Limitations are what kept the expectations from being met
within the context of this section.
Several scope changes were made to allow for the project to be completed keeping
consideration of researcher abilities and time constraints. The initial intention was to evaluate all
roadways within Kentucky (state or locally maintained) because of the nature of collisions.
Collisions occur regardless of road ownership, and reducing the type of roadways evaluated
could lead to segments being missed in dire need of attention by transportation agencies. For this
reason, the hot spot analysis was completed on all motorcycle involved collision records to
ensure the hot spot analysis layer was a valid representation of the areas experiencing a
statistically significant amount of collisions. The technical work for identifying suitable roadway
segments, however, was restricted to state-maintained roadways because of data availability. The

50

methodology for identifying suitable roadway segments can be applied to a roadway regardless
of location of ownership.
On the subject of data availability, a further limitation to this project was the lack of
recorded motorcycle AADT. The AADT values used within this project were for all motorized
vehicles – it was not specific to motorcycles although motorcycle involved collisions were the
only collisions evaluated. When looking at the nature of a motorcycle exclusive lane, there needs
to be a population present to use the lane as a validation component for modifying the
infrastructure. As mentioned before, motorcycle involved collisions were the only collisions
evaluated in lieu of having motorcycle AADT but having a motorcycle specific AADT data layer
would have facilitated demonstrating the need in an area for an exclusive motorcycle lane.
Research in Florida on motorcycle involved collisions and modeling future collision locations
has shown mixed results on the benefits of having motorcycle specific AADT - rural and urban
arterials showed an improved prediction of motorcycle involved collisions when using
motorcycle AADT but predictions on rural and urban freeways showed either negligible or
reduced prediction abilities when using motorcycle AADT (Lyon et al. 2016, 114).
With regard to the methodology created for accomplishing this project’s objective, the
workflow is replicable, but it could not be completely automated because of variations specific to
the data and study areas. The application of the methodology to multiple study areas helps
validate the ability to apply it to locations outside of Kentucky and factor in different variables
(i.e., using the percentage of curve for a roadway as opposed to the AADT value or widening
potential). The complete process, however, cannot be replicated by a model without custom
scripting. The first issue arose with the pixel size used in the Optimized Hot Spot Analysis tool.
Trial and error were the methods used in selecting the “valid” and final hot spot layer. When

51

running the tool without specifying a pixel size, the tool takes into account the record count and
polygon study area to calculate a default pixel size and this default changes for every layer
evaluated.
The varied pixel counts may not be an issue for some types of analysis, but when trying
to select impacted roadways for collision counts, the pixel size needed to be smaller in every run
of the model. An array of pixel sizes was selected for the study areas including 100m, 250m, and
500m and z (the default obtained by not entering a pixel size). Refer to Figure 25 and Figure 26
to see an example of Jefferson County with the default pixel size and the pixel size set at 250m.
The OHSA tool would be used for each of the layers at those pixel values - unless a visual
review of the resulting layer indicated the tool was making every record its own hot spot (refer to
figure 27 to see this happening in Meade County – these are not actual hotspots, they are every
point within the layer being evaluated). Study areas with a sufficient number of records to run
Figure 25. Jefferson County 250 m cell Figure 26. Jefferson County default cell

52

the OHSA tool but an insufficient amount to
detect hot spots were producing OHSA output
layers that incorrectly represented hot spots. A
manual review of every OHSA was needed to
determine if the selected pixel size was
appropriate and this impacted the ability to
replicate the methodology as a model for the
OHSA portion. A potential model could be
created using scripting languages such as Python
to compare the number of identified hotspots
with the number of records in the layer being
evaluated but that solution could not be explored
with the time constraints of this project. Additionally, several layers within Livingston and
Meade county failed because they did not have the number of records required to run the OHSA
tool (30 records) – this could also be accounted for within a custom script but it fell outside the
scope of this project.
5.2. Future Work
The work completed within this project can be applied to different areas, for different mitigation
measures, and with different collision factors. The steps within the methodology can be
replicated for different study areas as long as the data layers used for this project have an
equivalent layer for the new study area. It is important to clarify or reiterate this project
evaluated state-maintained roadways for suitable locations to implement a motorcycle lane, using
the relationship between higher traffic flow volumes with an increased probability of motorcycle
Figure 27. Meade County example

53

involved collisions. The AADT data used within this project did not have a motorcycle-specific
AADT value as the Kentucky Transportation Cabinet does not track that information – a future
project could use an area’s recorded motorcycle AADT if it is known to refine the suitable
locations from those obtained through an all-vehicle AADT value. Future work for this project
could also include evaluating additional collision factors including curve, slopes, speed limits,
pavement roughness, etc. and adjusting the safety measure being evaluated to reflect the collision
factor being evaluated. For instance, the collision factor of shoulder type (i.e. rumble strip,
concrete, guardrail) could be evaluated to determine if the shoulder type in an area is
contributing to the frequency or severity of an accident.
Future avenues this research could take also include the evaluation of motorcycle
involved collisions at intersections and in areas with an increased truck AADT, and the
integration of geodesign with results from the project. Focus for future work stemming from the
results of this project could center on motorcycle involved collisions specifically near or at
intersections. The existing infrastructure (signs) and line of sigh could also be evaluated to
determine if infrastructure changes could reduce the occurrence of motorcycle involved
collisions. This specific facet was not analyzed within this project because the project objective
dealt with higher flow areas and the infrastructure change for a motorcycle lane but if an area
had collision records, a layer of intersections or nodes, and the infrastructure surrounding
intersections, similar analysis could be completed. Similar analysis could also be performed
using semi-truck AADT to determine if motorcycle involved collisions occur more frequently in
an area with a higher truck presence. This could potentially indicate line of sight limitations
stemming from the size of semi-trucks as a semi-truck could impact the ability of other motorists
to see a motorcyclist.

54

In this regard, the topic of geodesign could also be explored because the findings from
these studies can be utilized in designing a road with geographic components in mind to address
factors as they are identified. Geodesign – within the context of this paper – is an emerging area
of work that unites the ability to plan, predict, and interact with geographic features in one focus
(Ervin 2012). Using predictive models or analysis completed on existing data as a base and
attempting to create solutions for phenomena being experienced (such as an increased occurrence
of motorcycle involved collisions in a certain area) is a possible application of geodesign. In
future work, the application of geodesign can be further explored in addressing identified hotspot
locations with various factors present to create a solution which minimizes the impact of those
factors within a geographic context.
Additional collision attributes could also be evaluated in future work revolving around
increasing motorcycle safety. For instance, if age or rider experience is recorded with collision
data, collision records could be evaluated to see if younger or inexperienced riders in a certain
region have collisions on a more frequent basis. If this is the case, that area could be targeted by
transportation agencies with a safety campaign for motorcycle classes aimed at increasing rider
maneuvering ability and education. This project helped create a base with a working
methodology that unites the existing identification of collision factors with the application of
mitigation measures for increasing motorcycle safety. The potential future work revolving
around GIS and increasing motorcycle safety is limited only by the data available and
willingness to complete the project.

55

References
"10 Common Motorcycle Crashes and How to Avoid Them." RideApart. Last modified
November 15, 2018. https://www.rideapart.com/articles/254912/10-common-motorcycle-
accidents-and-how-to-avoid-them/

“Fatality Facts 2017 Motorcycles and ATVs.” IIHS HLDI. Last modified December 2018.
https://www.iihs.org/topics/fatality-statistics/detail/motorcycles-and-atvs#fn1ref1

“Motorcycle Safety.” USDOT NHTSA. Last modified 2019. https://www.nhtsa.gov/road-
safety/motorcycle-safety

“Motorcycle Safety Campaign – Share the Road.” Texas Department of Transportation. Last
modified 2019. https://www.txdot.gov/driver/share-road/share-road.html

Allen, Trevor, Newstead, S., Lenné, M.G., Mcclure, R., Hillard, P., Symmons, M., and Day, L.
“Contributing Factors to Motorcycle Injury Crashes in Victoria, Australia.”
Transportation Research Part F: Psychology and Behaviour 45 (February 2017): 157–
168.

BikeBandit. “The Best Mobile Apps for Motorcycle Riders.” BikeBandit. Last Modified May 4,
2018. https://www.bikebandit.com/blog/the-best-mobile-apps-for-motorcycle-riders

Cheng, Andy S.K., Ng, Terry C.K., and Lee, Hoe C. “A Comparison of the Hazard Perception
Ability of Accident-Involved and Accident-Free Motorcycle Riders.” Accident Analysis
and Prevention 43, no. 4 (2011): 1464–1471.

Daniello, Allison, Kimberly Swanseen, Yusuf A. Mehta, and Hampton C. Gabler. “Rating Roads
for Motorcyclist Safety: Development of a Motorcycle Road Assessment Program.”
Transportation Research Record 2194, no. 1 (January 2010): 67–74. doi:10.3141/2194-
08.
Department of Geographic Information. (July 2019). Kentucky Geography Network.
https://kygeoportal.ky.gov/geoportal/catalog/search/search.page

Ervin, Stephen M. "Geodesign Futures–Nearly 50 predictions." In Speech at DLA 2012
Conference: Geodesign, 3D Modeling, and Visualisation. 2012.

Esri. “Optimized Hot spot Analysis.” Last modified 2019. https://pro.arcgis.com/en/pro-app/tool-
reference/spatial-statistics/optimized-hot-spot-analysis.htm

Gales, Morgan. "6 Smartphone Apps for Motorcyclists." MotorcycleCruiser. Last Modified
January 13, 2017. https://www.motorcyclecruiser.com/6-smartphone-apps-for-
motorcyclists/

56

Guido. “The Best Mobile Apps for Ride Planning, GPS Navigation and Tracking on iPhone –
2017 edition.” MotoMappers. Last Modified February 24, 2017.
http://www.motomappers.com/blog/best-motorcycle-apps-gps-navigation-ride-tracking/

Kentucky Office of Highway Safety. “2018 Highway Safety Performance Plan.” Kentucky
Transportation Cabinet. (September 30, 2018)

Kentucky State Police. “Kentucky Collision Analysis for the Public.” Last modified 2019.
http://crashinformationky.org/

Kentucky Transportation Cabinet. DataMart. (July 2019).
http://datamart.business.transportation.ky.gov/index.html

Kentucky Transportation Cabinet. “Standards for Road Data Collection & Maintenance Using
Global Positioning System Techniques.” (June 30, 2004).

Law, Michael, and Amy Collins. Getting to know ArcGIS PRO. Esri press, 2016.

Law, Teik Hua, Radin Umar Radin Sohadi, “Determination of Comfortable Safe Width in an
Exclusive Motorcycle Lane”, Journal of the Eastern Asia Society for Transportation
Studies, 2005, p. 3372-3385, Online ISSN 1881-
1124, https://doi.org/10.11175/easts.6.3372

Le, To Quyen, and Zuni Asih Nurhidayati. "A Study of Motorcycle Lane Design in Some Asian
Countries." Procedia engineering 142 (2016): 292-298.

Lyon, Craig, Bhagwant Persaud, Robert A. Scopatz, Scott Himes, Matt Albee, and Thanh Lee.
Investigating the Impact of Lack of Motorcycle Annual Average Daily Traffic Data in
Crash Modeling and the Estimation of Crash Modification Factors. No. FHWA-HRT-16-
054. United States. Federal Highway Administration, 2016.

Murthy, Sudhir, and Sinha, Kumares C. “A Fuzzy Set Approach for Bridge Traffic Safety
Evaluation 1.” Civil Engineering Systems 7, no. 1 (March 1, 1990): 36–43.

Nabors, Dan, Elissa Goughnour and Jon Soika. “Motorcycle Road Safety Audit Case Studies.”
National Technical Information Service, no. FHWA-SA-16-026 (May 2016).

Neira, Jorge, and Laura Bosque. 2004. “The Word ‘Accident’: No Chance, No Error, No
Destiny.” Prehospital and Disaster Medicine 19 (3). Cambridge University Press: 188–
89. doi:10.1017/S1049023X0000176X.

Nicol, David A., Dennis Heuwe and Dr. Susan Chrysler. “Infrastructure Countermeasures to
Mitigate Motorcyclist Crashes in Europe.” American Trade Initiatives, no. FHWA-PL-
12-028 (August 2012).

57

Radin Sohadi, Radin Umar, Mackay, Murray, and Hills, Brian. “Multivariate Analysis of
Motorcycle Accidents and the Effects of Exclusive Motorcycle Lanes in Malaysia.”
Journal of Crash Prevention and Injury Control 2, no. 1 (March 1, 2000): 11–17.

Ramirez, Philip Peter, Derek Otten, Dana Ferguson, Julie A. Jordan Peters, Regina Madigan, and
Thomas Mckenna. "Road segment safety rating system." U.S. Patent Application
15/015,623, filed June 16, 2016.

Rezapour, Mahdi, Moomen, Milhan, and Ksaibati, Khaled. “Ordered Logistic Models of
Influencing Factors on Crash Injury Severity of Single and Multiple-Vehicle Downgrade
Crashes: A Case Study in Wyoming.” Journal of safety research 68 (February 2019):
107–118.

Rice, Thomas, Lara Troszack and Taryn Erhardt. “Motorcycle Lane Splitting and Safety in
California.” Safe Transportation Research & Education Center. (May 29, 2015).

Richard, C. M., Magee, K., Bacon-Abdelmoteleb, P., & Brown, J. L. (2018, April).
Countermeasures that work: A highway safety countermeasure guide for State Highway
Safety Offices, Ninth edition (Report No. DOT HS 812 478). Washington, DC: National
Highway Traffic Safety Administration.

de Leur, Paul and Sayed, T. “Development of a Road Safety Risk Index.” Statistical
Methodology: Applications To Design, Data Analysis, And Evaluat, no. 1784 (2002):
33–42.

Schneider, William H., Savolainen, Peter T., Van Boxel, Dan, and Beverley, Rick. “Examination
of Factors Determining Fault in Two-Vehicle Motorcycle Crashes.” Accident Analysis
and Prevention 45 (March 2012): 669–676.

Sultan, Vivian, Ahmed Alzahrani, Hind Bitar, and Norah Alharbi. "Is California's Aging
Infrastructure the Principal Contributor to the Recent Trend of Power Outage?." Journal
of Communication and Computer 13 (2016): 225-233.

Zahranel-Said M. M, Soon Jiann Tan, ’asri Putra Nurul Amirah ’atiqah Binti Mohamad, Tan Eng
Hie Angel, Yap Yok Hoe, and Rahman Ena Kartina Abdul. “Evaluation of Various GIS-
Based Methods for the Analysis of Road Traffic Accident Hotspot.” MATEC Web of
Conferences 258 (January 1, 2019).
https://doaj.org/article/6f43c95d36db46b1ab55f4a8af93ac04

58

Appendix A: Queries

Study Area Motivation Category Focus Area Definition Query
Multiple Vehicles UnitsInvolved <> 1 Or UnitsInvolved <> 2
Single Vehicle UnitsInvolved = 2
Motorcycle Only UnitsInvolved = 1
No Casualty NumberKilled = 0 And NumberInjured = 0
Fatality Greater than or equal to 1 NumberKilled <> 0
Injury greater than or equal to 1 NumberInjured <> 0
Master All Collisions
Multiple Vehicles UnitsInvolved <> 1 And UnitsInvolved <> 2 And County = 'MEADE'
Single Vehicle UnitsInvolved = 2 And County = 'MEADE'
Motorcycle Only UnitsInvolved = 1 And County = 'MEADE'
No Casualty NumberKilled = 0 And NumberInjured = 0 And County = 'MEADE'
Fatality Greater than or equal to 1 NumberKilled <> 0 And County = 'MEADE'
Injury greater than or equal to 1 NumberInjured <> 0 And County = 'MEADE'
Master All County Collisions County = 'MEADE'
Multiple Vehicles UnitsInvolved <> 1 And UnitsInvolved <> 2 And County = 'JEFFERSON'
Single Vehicle UnitsInvolved = 2 And County = 'JEFFERSON'
Motorcycle Only UnitsInvolved = 1 And County = 'JEFFERSON'
No Casualty NumberKilled = 0 And NumberInjured = 0 And County = 'JEFFERSON'
Fatality Greater than or equal to 1 NumberKilled <> 0 And County = 'JEFFERSON'
Injury greater than or equal to 1 NumberInjured <> 0 And County = 'JEFFERSON'
Master All County Collisions County = 'JEFFERSON'
Multiple Vehicles UnitsInvolved <> 1 And UnitsInvolved <> 2 And County = 'LIVINGSTON'
Single Vehicle UnitsInvolved = 2 And County = 'LIVINGSTON'
Motorcycle Only UnitsInvolved = 1 And County = 'LIVINGSTON'
No Casualty NumberKilled = 0 And NumberInjured = 0 And County = 'LIVINGSTON'
Fatality Greater than or equal to 1 NumberKilled <> 0 And County = 'LIVINGSTON'
Injury greater than or equal to 1 NumberInjured <> 0 And County = 'LIVINGSTON'
Master All County Collisions County = 'LIVINGSTON'
Multiple Vehicles UnitsInvolved <> 1 And UnitsInvolved <> 2 And County = 'FAYETTE'
Single Vehicle UnitsInvolved = 2 And County = 'FAYETTE'
Motorcycle Only UnitsInvolved = 1 And County = 'FAYETTE'
No Casualty NumberKilled = 0 And NumberInjured = 0 And County = 'FAYETTE'
Fatality Greater than or equal to 1 NumberKilled <> 0 And County = 'FAYETTE'
Injury greater than or equal to 1 NumberInjured <> 0 And County = 'FAYETTE'
Master All County Collisions County = 'FAYETTE'
Injury Status
Vehicle Status
Injury Status
Vehicle Status
Fayette
County, KY
Second
highest
cumulative
total of
motorcycle
involved
collisions
Vehicle Status
Injury Status
Kentucky
Meade
County, KY
Jefferson
County, KY
Livingston
County, KY
Higher
percentage
of
motorcycle
involved
collisions
resulting in
Highest
cumulative
amount of
motorcycle
involved
collisions by
county
Higher
percentage
of
motorcycle
involved
collisions
resulting in
Statewide
comparrison
- acts as a
control for
the specific,
county
based
Injury Status
Vehicle Status
Injury Status
Vehicle Status

59

Appendix B: Results

Study Area Motivation Category # of Records Status 100m 250m 500m 5000m **Z m Results
8948 Complete NC NC NC CI NC
9352 Complete NC NC NC CI NC
8322 Complete NC NC NC CI NC
6260 Complete NC NC NC CI NC
842 Complete NC NC NC CI NC
11408 Complete NC NC NC CI NC
Master 18300 Complete NC NC NC CI NC
3 Failed NC NC NC NC NC *ERROR 001570
44 Complete NC CV CI NC NC
61 Complete NC NC CV NC NC
25 Failed NC NC NC NC NC *ERROR 001570
16 Failed NC NC NC NC NC *ERROR 001570
71 Complete NC NC CV NC NC
Master 108 Complete CI CI CV NC NC
157 Complete CI CV CV NC CI
2099 Complete CI CV CI NC CI
904 Complete CV CV CI NC CI
1109 Complete NC CV NC NC CI
141 Complete NC CV CV NC CI
1951 Complete CV CV NC NC CI
Master 3160 Complete CV CV CI NC CI
1 Failed NC NC NC NC NC *ERROR 001570
17 Failed NC NC NC NC NC *ERROR 001570
39 Complete NC NC CV NC CI
10 Failed NC NC NC NC NC *ERROR 001570
1 Failed NC NC NC NC NC *ERROR 001570
46 Complete NC NC CV NC CI
Master 57 Complete NC NC CV NC CI
50 Complete NC CI CV NC CI
837 Complete NC CV NC NC CI
384 Complete NC CV NC NC CI
506 Complete NC CV CI NC CI
36 Complete NC CI CI NC CI
735 Complete NC CV NC NC CI
Master 1271 Complete CV CI CI NC CI
*ERROR 001570 The analysis option you selected requires a minimum of 30 points to be inside the bounding polygon area(s)
**Z m Tool was completed using automatically generated cell size
CV Completed, Viable Result
NC Not Completed
CI Completed, Inconclusive
Injury Status
Vehicle Status
Injury Status
Vehicle Status
Fayette
County, KY
Second highest
cumulative total of
motorcycle involved
collisions resulting in
fatality and injury
Vehicle Status
Injury Status
Kentucky
Meade
County, KY
Jefferson
County, KY
Livingston
County, KY
Higher percentage of
motorcycle involved
collisions resulting in
injuries by county
total
Highest cumulative
amount of
motorcycle involved
collisions by county
Higher percentage of
motorcycle involved
collisions resulting in
fatalities by county
Statewide
comparrison - acts as
a control for the
specific, county
based analysis
Injury Status
Vehicle Status
Injury Status
Vehicle Status

60

Appendix C: Final Results

Study Area Motivation Focus Area # of Records Status Final Results
Meade
County, KY
Higher percentage of
motorcycle involved collisions
resulting in fatalities by county
All County Collisions 108 Incomplete No viable locations identified -
county does not contain state-
maintained roads meeting or
exceeding 60,000 AADT.
Jefferson
County, KY
Highest cumulative amount of
motorcycle involved collisions
by county
All County Collisions 3160 Complete Multiple locations with need
and infrastructure capable of
supporting a motorcycle lane
identified.
Livingston
County, KY
Higher percentage of
motorcycle involved collisions
resulting in injuries by county
total
All County Collisions 57 Incomplete No viable locations identified -
county does not contain state-
maintained roads meeting or
exceeding 60,000 AADT
intersecting the identified hot
spots.
Fayette
County, KY
Second highest cumulative
total of motorcycle involved
collisions resulting in fatality
and injury
All County Collisions 1271 Complete Multiple locations with need
and infrastructure capable of
supporting a motorcycle lane
identified.

61

Appendix D: Commonwealth of Kentucky

62

63

64

65

Appendix E: Livingston County

66

67

68

Appendix F: Meade County

69

70

71

72

Appendix G: Jefferson County

73

74

75

76

77

78

79

80

Appendix H: Fayette County

81

82

83

84

85

86

Abstract (if available)

Abstract Motorcycles are disproportionately affected in collisions when compared to other motor vehicle types, leading to an increased vulnerability of injury or death to motorcyclists. Multiple factors can contribute to this disproportionate impact, including environmental factors, inattention of other motorists, driver error, and physical road characteristics. Many motorcycle safety initiatives address the error and role of other drivers in motorcycle-involved collisions but little attention is often given to the environmental and roadway factors. This lack of attention and analysis reduces the ability of transportation agencies to obtain a complete common operating picture for all factors impacting motorcycle safety – allowing for missed opportunities to decrease the increased vulnerability of motorcyclists. ❧ This project utilized Geographic Information Systems (GIS) as a tool to identify locations on state-maintained roadways showing statistically significant clusters of motorcycle involved collisions. The collision data for this report were retrieved from the Kentucky State Police collision database

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Alaska Hike Search: designing a mapping application for Alaskan trails and user contributed hazards

PDF

Developing art-based cultural experiences in North Kohala: A community engagement project with OneIsland

PDF

Building a spatial database of biochar research and practice with Web-GIS

PDF

Coastal vulnerability assessment for archaeological sites on San Clemente Island and San Nicolas Island, California

PDF

Analysis of park accessibility in Redan, Georgia Web GIS application

PDF

Congestion effects on arterials as a result of incidents on nearby freeway: When should you get off the highway?

PDF

Creating a Web GIS to support field operations and enhance data collection for the Animal and Plant Health Inspection Service (APHIS)

PDF

A model for placement of modular pump storage hydroelectricity systems

PDF

Site suitability analysis for implementing tidal energy technology in southern California

PDF

A spatiotemporal analysis of racial disparity in the distribution of superfund sites within Santa Clara County, California

PDF

Suitable California Opportunity Zone (COZ) locations for affordable housing development in the cities of Bakersfield, Los Angeles, and Palmdale

PDF

Wrong way driving in San Antonio, Texas: a transportation route study using Network Analyst

PDF

Creating a water quality geodatabase for the West Hawai‘i Island region

PDF

Building a geodatabase design for American Pika presence and absence data

PDF

Cal ToxTrack: a full stack Web GIS for mapping pollution in California

PDF

A spatiotemporal analysis of environmental risk factors of Lyme disease in the Northeastern United States

PDF

Geographic object based image analysis for utility scale photovoltaic site suitability studies

PDF

Designing an earthquake preparedness web mapping application for the older adult population of Los Angeles, California

PDF

Suitability analysis for wave energy farms off the coast of Southern California: an integrated site selection methodology

PDF

Access to active play parks for youth segments in Alexandria, Virginia

Asset Metadata

Creator Bartee, Emily Carol (author)

Core Title Roadway hazard analysis: a safe ride for motorcycles

School College of Letters, Arts and Sciences

Degree Master of Science

Degree Program Geographic Information Science and Technology

Publication Date 02/07/2020

Defense Date 12/18/2019

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag analysis, hotspot analysis,ArcGIS Pro,average annual daily traffic,collisions,geodesign,geographic information systems,GIS,hazard analysis,Kentucky,lane widening,motorcycle accidents,motorcycle collisions,motorcycle lane,motorcycle safety,OAI-PMH Harvest,optimized hotspot analysis,Transportation

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Bernstein, Jennifer (committee chair), Duan, Leilei (committee member), Swift, Jennifer (committee member)

Creator Email ebartee@usc.edu,emily.c.bartee@gmail.com

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c89-266023

Unique identifier UC11673068

Identifier etd-BarteeEmil-8155.pdf (filename),usctheses-c89-266023 (legacy record id)

Legacy Identifier etd-BarteeEmil-8155.pdf

Dmrecord 266023

Document Type Thesis

Rights Bartee, Emily Carol

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the a...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA