DEVELOPMENT OF A GIS-BASED TRAFFIC SAFETY ANALYSIS SYSTEM 1 2 Vidhya Kumaresan, M.S.E., E.I.T. 3 Research Associate 4 5 Vinod Vasudevan, M.S.E., P.E. 6 Associate Research Engineer 7 8 Transportation Research Center 9 University of Nevada, Las Vegas 10 4505 S. Maryland Parkway, Box 454007 11 Las Vegas, NV 89154-4007 USA 12 Tel.: (702) 895-1393 13 E-mail: [email protected]; [email protected]; 14 15 16 Shashi S. Nambisan, Ph.D., P.E. 17 Director, Institute for Transportation 18 Professor of Civil Engineering 19 Iowa State University 20 2711 South Loop Drive, Suite 4700, Ames, IA 50010-8664 USA 21 Tel: (515) 294-5209, E-mail: [email protected]22 23 Jerry Duke 24 Manager of Planning 25 Regional Transportation Commission of Southern Nevada 26 600 S. Grand Central Parkway 27 Las Vegas, NV 89106, USA 28 Tel (702) 676-1729, E-mail: [email protected]29 30 31 32 33 34 35 Word Count: Text (3,877+ Figures/Tables 13 × 250) = 7,127 total words 36 37 38 39 40 41 42 43 Submitted for publication and presentation at the 44 89 th Annual Transportation Research Board Meeting 45
Transcript
DEVELOPMENT OF A GIS-BASED TRAFFIC SAFETY ANALYSIS SYSTEM 1 2
Vidhya Kumaresan, M.S.E., E.I.T. 3 Research Associate 4
5 Vinod Vasudevan, M.S.E., P.E. 6 Associate Research Engineer 7
8 Transportation Research Center 9
University of Nevada, Las Vegas 10 4505 S. Maryland Parkway, Box 454007 11
Las Vegas, NV 89154-4007 USA 12 Tel.: (702) 895-1393 13
Word Count: Text (3,877+ Figures/Tables 13 × 250) = 7,127 total words 36 37 38 39 40 41 42 43
Submitted for publication and presentation at the 44 89th Annual Transportation Research Board Meeting45
Kumaresan, Vasudevan, Nambisan, and Duke
1
Development of a GIS-based Traffic Safety Analysis System 1 Vidhya Kumaresan1, Vinod Vasudevan2, Shashi S. Nambisan3 and Jerry Duke4 2
3 4
ABSTRACT 5
Traffic safety analyses have historically been used by transportation professionals as a means to 6 evaluate safety of an area. This paper summarizes the methodology to develop a traffic safety 7 analysis system – one that integrates crash data and roadway related information in a Geographic 8 Information System (GIS) environment. The system includes customized user interfaces to 9 support queries and capabilities to export results. This system supports both macro level and 10 micro level analyses. The system further offers capabilities to query crash data based on 11 attributes from the database. The system also contains a module to identify high crash locations 12 based on methods identified from the published literature. The methods range from those based 13 on simple crash frequency to more complex methods which incorporate different weights for 14 crashed based on the crash outcomes. The system can be used to evaluate the effectiveness of a 15 safety program by performing ‘before and after’ studies. An application of the system is 16 illustrated using data from the Las Vegas metropolitan area in the state of Nevada. 17 18 INTRODUCTION 19
Over the past decade, traffic crashes have resulted in over 42,000 fatalities annually in the United 20 States (1). Analyses of relevant data are essential to develop strategies to improve road safety. 21 However, data required for such analyses often come from multiple sources, in varied formats, 22 with various levels of accuracy and reliability. Computerized tools and systems offer great 23 potential to combine such data and to perform the required analyses. Due to the spatial nature of 24 disparate datasets needed for transportation safety analysis, the GIS platform would be a good 25 option to integrate them. Kumaresan provides additional information on the advantages of using 26 GIS for development of such a system (2). Further, the GIS platform offers a natural 27 environment for development of a safety analysis system using the consolidated database due to 28 its data integration and mapping capabilities. The advantage of viewing data in spatial format, 29 hence enabling the user for spatial analyses, is perhaps a very important benefit of using GIS. 30 31
32
1 Research Assistant, Transportation Research Center, University of Nevada, Las Vegas, 4505 S.
2 Associate Research Engineer and Program Manager, Transportation Research Center, University of Nevada, Las Vegas, 4505 S. Maryland Parkway, Box 454007, Las Vegas, NV 89154-4007; Tel.: (702)895-1594; E-mail: [email protected]
3 Professor of Civil Engineering and Director, Institute for Transportation, Iowa State University, 2711 South Loop Drive, Suite 4700, Ames, IA 50010-8664; Tel: 515-294-5209; E-mail: [email protected]
4 Manager of Planning, Regional Transportation Commission of Southern Nevada, 600 S. Grand Central Parkway, Las Vegas, NV 89106; Tel: 702-676-1729; E-mail: [email protected]
Kumaresan, Vasudevan, Nambisan, and Duke
2
LITERATURE REVIEW 1 Analysis of Crash Data 2 Several efforts have been initiated throughout United States to create crash reporting systems 3 which allow agencies to access these data. The Fatality Analysis Reporting System (FARS) 4 developed by the National Highway Traffic Safety Administration (NHTSA) (3) is one such 5 system that aims to provide a measure of overall highway safety and thereby to evaluate the 6 effectiveness of highway safety programs. Parrish et. al. (2003) explains the Critical Analysis 7 Reporting Environment (CARE), developed by the University of Alabama (4). The aim of CARE 8 was to analyze automobile crashes to aid in developing strategies to reduce crashes. Similarly, 9 University of North Carolina’s Highway Safety Research Center (HSRC) has developed a crash 10 data query website that’s contains data from 2001-2006 available for public usage Alabama (5). 11 Iowa Department of Transportation uses Crash Mapping Analysis Tool (CMAT) to provide 12 access to crash data and additional features like crash stacking, inclusion of speed and volume 13 information (6). Kim et al. describes the Crash Outcome Data Evaluation System (CODES) 14 system used in Hawaii, New Hampshire, Maryland and South Carolina (7). CODES involves the 15 linkage and analysis of police crash data, emergency medical services transport data, hospital 16 data, and insurance claims. Many state agencies like Nevada Department of Transportation (8), 17 Washington State Department of Transportation (9) and Department of Transportation of 18 Arlington (10) have developed systems to aid in traffic safety analysis and management. 19 20 Advantages of Using GIS to Develop Crash Analysis Tool 21 Motor vehicle crash data for metropolitan regions are typically collected and maintained by 22 several agencies such as federal, state, local agencies and law enforcement agencies. Therefore, 23 there likely are significant differences in their formats and in different data structures. In order to 24 integrate crash data with other relevant information like roadway data, GIS provides an 25 environment to facilitate data integration, analysis (11) and display of results. The use of GIS 26 also extends the level of analyses to perform spatial analyses and grid-cell modeling (12). More 27 detailed commentary on the rationale for this approach is included in a report that forms the basis 28 of this paper (13). 29 30 OBJECTIVE 31 The literature review suggested that although some studies are conducted in analyzing crash data 32 using GIS, these features have to be tailored to fit agencies’ analyses requirement. Implementing 33 a bigger system is both time consuming and expensive. On the other hand, most of the existing 34 tools did not address micro analyses in detail. From a safety engineers’ and planners’ 35 perspective, it is extremely important to have a safety analysis system capable of performing 36 detailed micro level analyses for evaluating effectiveness of their safety programs. The objective 37 of this paper is to demonstrate the development of a safety analysis system using crash data and 38 roadway data. The aim of this paper is to explain the concepts of both macro and micro- analyses 39 for a region using the existing crash data and street centerline data. Examples from the Las 40 Vegas Metropolitan area are used to illustrate the capabilities of the system developed. 41 42 METHODOLOGY AND SYSTEM DEVELOPMENT 43
Data Assimilation 44 The initial step in the development of an analysis system is to identify the data required to 45 support the desired analyses and the format in which they are available. Data assimilated 46
Kumaresan, Vasudevan, Nambisan, and Duke
3
included crash data for the time period 2002 to 2006 from the Nevada Department of 1 Transportation and roadway network (street centerline) information from the Clark County GIS 2 Management Office (GISMO) (14). The crash database contains several crash characteristics 3 including vehicle related, roadway related, crash related and person related details (15). One of 4 the issues related to data assimilation in this research was the compatibility of both the data sets 5 after incorporating them in a GIS environment. 6
7 System Architecture and Design 8 The system is intended to support performing safety analyses in terms of querying crash data and 9 ranking of high crash locations. The analyses can be broadly classified into micro and macro 10 levels of examination. Macro level analyses involve analyses on a larger spatial scale such as a 11 city, while micro analyses focus at a smaller spatial extent such as an intersection or roadway 12 segment. As for output, the system was required to support representation of the results in a form 13 that would be acceptable for most users. Furthermore, the software and hardware requirements 14 were to be easily available to the users. 15 16 METHODOLOGY AND SYSTEM DEVELOPMENT 17
System Development 18 The safety analysis system was developed for the crash data as a whole and another section was 19 dedicated solely to bicycle and pedestrian crashes for analyzing and evaluating improvements to 20 those particular categories of road users. The development of the entire safety analysis system 21 can be described in three parts: 1) macro-analyses; 2) micro-analyses – intersection and street 22 segment; and 3) the high crash ranking analysis system for various locations. Figures 1a, 1b, 23 and1c describe the tool structure of the above-mentioned sections. 24 Macro-Analyses 25 The system provided the user with options to query for crash data on a macro (area-wide) level 26 or on micro (location-specific) level. After deciding on the type of query the system listed the 27 options that could be included in querying the crash data pertaining to the specific location. Most 28 of the attributes listed in the crash database such as severity, crash type, day of week, time of day 29 etc. could be included in developing a query. A basic search option is provided to the user to 30 perform a quick search of the crash data for any criterion that the database supported on a macro 31 or area wide level. The user entry in this section is restricted to three entries. This option 32 facilitates the user to work with a combination of 3 criteria of crash characteristics. A sample of 33 this application is shown in the case study section. 34 Micro-Analyses 35 Crash analyses in micro-level can be performed for locations – either intersections or street 36 segments. As Figure 1b indicates, analyses based on intersection or segment would allow the 37 user to further make use of multiple combinations of queries to arrive at specific results. The 38 following section describes both these options in detail. 39 (a) Crashes at an Intersection - The main objective of the intersection level analysis option is to 40
analyze safety aspects at an intersection location. The data entry section involved the entry of 41 the names of the cross streets and the buffer distance to be considered around the intersection 42 fr which crash data is to be extracted. The flow of the logic is represented in figure 2. 43
(b) Crashes on a Street Segment - The third option for crash analysis is based on identifying 44 street segments. This option permits evaluation of long segment of a street comprised of 45 several smaller sections or just one section. The data entry for the street segment analyses 46
Kumaresan, Vasudevan, Nambisan, and Duke
4
1 a Macro - Analysis Tool Structure. 2
3
4
5 b Micro-Analyses Tool Structure – intersection and street segment. 6
7
8 c High Crash Ranking Tool Structure. 9
10 FIGURE 1 Tool structure. 11
12
Data – crash, roadway
Crash frequency for areas
Assign ranks
Frequency Weights
Multiple Queries
Severity Crash type Time period Day of week
Time of day Roadway factors Vehicle Details Weather
FIGURE 2 Flow chart for selecting crashes for intersection level analysis. 32 33 34
required the choice of the names of the main street as well as the cross streets between which 35 the main street segment was to be located. The coding tasks were performed based on the 36 figure 3. 37
c) Multiple Query Criteria - Based on the selected intersection or street segment, the system 38 provides the user an interface to build a query based on various user defined criteria using 39 data fields from the crash database. The user options to select the criteria for the multiple 40 query analysis are placed on four tabs in four broad categories as follows: date & time, crash 41 characteristics, vehicle characteristics, and road characteristics. 42 43 44
FIGURE 3 Flow chart for selecting crashes for street segment based analysis. 40 41 42 High Crash Ranking 43 The system facilitates the identification of high crash locations based on total crashes and 44 weighted ranking (16). The methodology followed for identification of high crash ranking was 45 (a) simple crash frequency procedure – assigning ranks based on total number of crashes in that 46
Search
Select main street
Draw buffer around main street
Export streets within main street buffer
Select cross streets from extracted streets
Draw buffer to select crashes around main street
OK
No
Select main street segment from overlapped polygon
Main street Cross street 1 Cross street 2
Check data entry
Extract Crashes
Draw polygon using selected cross streets as sides
Kumaresan, Vasudevan, Nambisan, and Duke
7
particular location and (b) weighted ranking procedure – assigning ranks based on weights 1 specified for different levels of severity – fatal, injury or PDO. 2 3 CASE STUDY: LAS VEGAS METROPOLITAN AREA 4 The various analyses tools listed in the methodology section are illustrated using examples from 5 the Las Vegas metropolitan area in the case study section. ESRI’s ArcGIS software version 9.2 6 (17) was used to develop the safety analysis system. This was because of several advantages it 7 possesses in spatial display and data integration. Besides the technical capabilities it offered, this 8 was also the software used by agencies in Las Vegas Metropolitan area. Its in-built VBA 9 application was the programming environment used to develop the various analysis tools and 10 operations. Microsoft Excel was also used for displaying the analysis results in tabular format. 11 Another type of output possible using the ArcGIS software was to display the results graphically 12 on the map window. 13
14 Macro Analyses 15 This section describes the use of the tool in performing macro-level analyses. In order to 16 illustrate this option, the following sample query statement (broken down into three parts) is 17 shown: 18
First Criterion: Crash type – Head-on, and 19 Second Criterion: Severity – Injury, and 20 Third Criterion: Year – 2004 21 22
23 FIGURE 4 Final appearance of sample query in Simple Search window. 24
25 The first step for a user to perform this sample query was to click on the ‘First Criterion’ 26
combo-box (Figure 4, Label a). From the list, ‘Crash Type’ criterion had to be chosen. After 27 choosing ‘Crash Type’ as the data field to be queried on, the user had to click on the ‘Load 28 unique values’ button as shown in Figure 4, Label b. This would load all unique values in the 29 criterion to Figure 4, Label c. The next step was for the user to choose ‘Head on’ from the list. 30
a b c
d
Kumaresan, Vasudevan, Nambisan, and Duke
8
Thus, the first criterion is to set to crash type to “head on.” Next, the connection between the first 1 and second statement was to be mentioned (Figure 4, Label d). The operators available for use 2 are: ‘And’ and ‘Or’, and they would be displayed on clicking the down arrow button next to the 3 box. For this example, the use of the ‘And’ operator was required. 4
The same procedure of selection of criteria had to be followed for criteria two and three 5 also. The final appearance of the screen after data entry is shown in Figure 4. The results for this 6 query are shown in Figure 5 (zoomed in for clarity). 7
8
9
FIGURE 5 Results of the query: Angle crashes that caused injuries in 2005. 10 11 Micro Analyses 12 This section describes the performance of location-specific analyses in terms of whether it was 13 an intersection or a roadway segment. The tool for this option has been demonstrated using 14 suitable examples from the Las Vegas area. 15 a) Crashes at an Intersection 16 In the first step, the user had to choose the names of the intersecting streets. Clicking the button 17 shown in Figure 6, Label a resulted in the streets names of Las Vegas area from the Street 18 Centerline database being loaded on to the combo-box labeled ‘Choose the First Street.’ For 19 example, for the Rainbow-Sahara intersection, the user selected ‘Rainbow’ for the 1st street and 20 ‘Sahara’ for the 2nd street or vice versa. The next step was to enter the buffer distance into the 21 textbox titled ‘Input the Buffer Distance’ (Figure 6, Label b) and specify the unit (Figure 6, 22 Label c.) For illustration purposes, the buffer distance was chosen as 250 and the unit for this 23 distance was chosen as ‘feet.’ 24 Clicking the ‘Search’ button would prompt the system to check for the existence of the specified 25 intersection location and then performance of the sequence of operations to extract crashes that 26 occurred around the intersection. The output for the sample query that was run with the data 27 entry as given in the previous figure is shown in Figure 7. 28 29
FIGURE 7 Results: Crashes at Maryland Pkwy – Flamingo Rd for a 250 foot radius. 8 Note – There exists a minimal offset of the crash data from the street centerline database owing to the fact that the 9 crashes were geocoded to the street system while creating the database whereas the line represents the centerline of 10 the street and the offset corresponds to the original geocoding process. 11 12 b) Crashes on a Street Segment 13 In order to illustrate the operation extraction of crashes that occurred on a street segment, the 14 segment of Sahara Avenue between Buffalo Drive and Rainbow Boulevard is shown. The data 15 entry for this example and the results are shown in Figure 8. 16
17
a
bc
Kumaresan, Vasudevan, Nambisan, and Duke
10
1
a Street Segment query 2
3 4
b Results of crashes on Sahara between Buffalo and Rainbow 5 FIGURE 8 Segment Query and Results. 6
7 Multiple Query Criteria 8 After either of the two above-mentioned tools – location of crashes at an intersection or on a 9 street segment is performed; the multiple query criteria window pops open to enable the user to 10 include a variety of criteria in his/her query statement. For previously run street segment query 11 example, the multiple query window is shown in Figure 9. The various crash characteristics are 12 grouped under the 4 tabs highlighted in Figure 9, Label a: Date/Time, Crash, Vehicle, Road 13 details. Sample results for the query and export options (figure 10) are shown in Figure 11a 14 (shapefile) and 11b (table). 15
16
Kumaresan, Vasudevan, Nambisan, and Duke
11
1
FIGURE 9 Multiple Query Criteria screen. 2 3
4 5
FIGURE 10 Results of Multiple query criteria. 6 7 At this point the system allows the user to either export the results as a shapefile (Figure 11a) or 8 as a spreadsheet table (Figure 11b). 9 10
a
a b
Kumaresan, Vasudevan, Nambisan, and Duke
12
1 a Results of exported shapefile 2
3
4
b Results exported to Microsoft Excel document 5 6
FIGURE 11 Exporting Results from Query. 7 High Crash Ranking 8 In this section, the High Crash Ranking tool is described for a sample run in order to explain the 9 step by step procedure to be followed by the user to get required results with regard to whether it 10 was based on a) total number of crashes or b) weights for severity levels. 11 For the crash frequency procedure, assuming that the user wanted to rank the locations in Las 12 Vegas area based on the total number of crashes that occurred at each location in the year 2004 13 for a buffer distance of 200 feet around the intersection, the High Crash Ranking feature shown 14 in Figure 12, Label a had to be used. After clicking the High Crash Location button, the ranking 15 would be performed by the system. The resulting ranks obtained for the year 2004 and buffer 16 size 200 feet are shown in Figure 13a. If the user wishes to make use of the option to include 17 weights while ranking the locations, the feature shown in Figure 12, Label b had to be used. In 18 order illustrate this feature, assuming the user wished to append weights for severity to the 19 previously mentioned query year and buffer distance, fatal being 10, injury being 5 and PDO 20 being 2, clicking the ‘Weighted Ranking’ button triggered the system to calculate the sum of 21
Exported Crashes
Kumaresan, Vasudevan, Nambisan, and Duke
13
weights. The results obtained for the specified weights of severity levels having been 1 incorporated in the ranking procedure, are shown highlighted in Figure 13b. 2
3
4 5
FIGURE 12 High Crash Ranking screen 6 7
ISSUES AND ERRORS 8
The Safety Analysis System was developed using ArcGIS 9.2. However, after completion of the 9 project, ESRI released ArcGIS 9.3. When tried to use the tool in the new version, the system 10 showed some compatibility issues. These were resolved and the system was deployed 11 successfully. 12 CONCLUSIONS AND RECOMMENDATIONS 13
The main objective of this paper was to summarize the methodology to develop a safety analyses 14 system to perform analyses in both macro and micro level. The paper describes the development 15 of such a system using GIS software. This system could be used to identify safety issues in an 16 area. It could also be used to monitor traffic safety programs and their effectiveness. The safety 17 analysis system thus developed was able to perform crash analysis at a specific location in a 18 detailed manner as well as rank locations based on number of crashes. Some advantages of the 19 safety analysis system are: 20
The safety analysis system helps evaluate safety concerns on roadways in an area or at 21 specific locations. 22
The High Crash Ranking tool can be used to assign ranks to locations based on the total 23 number of crashes that occurred around it. 24
The safety analysis system enables users to integrate various formats of data and to obtain 25 results based on queries and spatial analyses and to display results as maps. 26
The safety analysis system can be used to perform before and after studies to determine 27 the effectiveness of a transportation safety program. Also, the user interface helps users 28 who are not familiar with GIS functionalities to make use of the developed analysis tools. 29
30 Some of the recommendations of the study are: 31
From the crash database, it was found that the records contained numerous misspelled 32 entries. Since the querying of data depended entirely on the individual entries in the 33
a b
Kumaresan, Vasudevan, Nambisan, and Duke
14
database, addressing all possible spellings of a particular factor could not be achieved 1 entirely. Creation of a reliable crash database would help to improve the accuracy of the 2 results obtained by using the various analysis tools. 3
Use of information technology by means of handheld/PDA type devices while recording 4 crash reports would help in minimizing data inconsistencies and misspells in addition to 5 the time saved in manual inclusion of each report into the crash database. 6
Incorporation of other traffic-related information such as traffic volumes/mix, traffic 7 control and operational information, roadway-related information in evaluating traffic 8 safety would help in improving the analyses possible. For example, computation of crash 9 rates such as the crashes per million entering vehicles at an intersection would provide a 10 measure for ranking crashes. 11
12 13
14 a Year 2004 and buffer size 200 feet. 15
16
17 18
b Weighted ranking for Severity – 10, 5 and 2. 19 20
FIGURE 13 High Crash Ranking results. 21 22
Kumaresan, Vasudevan, Nambisan, and Duke
15
1 2 ACKNOWLEDGMENTS 3
This paper is based significantly on a study sponsored by the Regional Transportation 4 Commission of Southern Nevada (RTC). It is also an outcome of the first author’s MSE thesis at 5 UNLV. The authors acknowledge the support provided by a number of individuals that made 6 possible the efforts documented in this manuscript. help Several representatives of the RTC, 7 particularly Mr. Shital Patel and Mr. Tom Wolch, provided valuable guidance to help identify 8 the system’s functionality. Ms. Xin Li from the UNLV Transportation Research Center and the 9 ESRI staff are thanked for their help in VBA programming. The crash data used in the analysis 10 were obtained from the Nevada Department of Transportation, and this assistance is gratefully 11 acknowledged. 12 13 REFERENCES 14
2. Kumaresan, V. 2008. Development of a GIS-based Safety Analysis System. M.S.E. Thesis. 17 University of Nevada Las Vegas. 18
3. Fatality Analysis Reporting System (FARS) Research and Development, http://www-19 fars.nhtsa.dot.gov/Main/index.aspx, July 28, 2009 20
4. Parrish, Allen S., Dixon, B., Cordes, D., Vrbsky, S. and Brown, D. June 2003. CARE: An 21 Automobile Crash Data Analysis Tool. IEEE Journal, 36(6), 22-30. 22
5. Highway Safety Research Center, University of North Carolina, 23 http://www.hsrc.unc.edu/index.cfm, July 28, 2009 24
6. Crash Mapping Analysis Tool (CMAT), Iowa Deaprtment of Transportation, 25 http://www.iowadot.gov/crashanalysis/cmatmain.htm, July 28, 2009 26
7. Kim, K., Kerns, T., Hettinger, T., and Pease, M. March 2001. National Highway Traffic 27 Safety Administration Report Number DOT HS 809 201, National Center for Statistics and 28 Analysis. 29
8. Safety Management System, Nevada Department of Transportation, 30 http://www.ndotsms.com/index.htm, July 28, 2009 31
9. Local Agency Safety Management, Washington Department of Transportation, 32 http://safety.fhwa.dot.gov/local_rural/fhwasa03025/, July 28, 2009 33
10. Miller, J.S. 2000. Geographical Information Systems: Unique Analytic Capabilities for the 34 Traffic Safety Community. Transportation Research Record 1734, 21-28. 35
11. Austin, K., Tight, M. and Kirby H. 1997. The Use of Geographical Information Systems to 36 Enhance Road Safety Analysis. Transportation Planning and Technology 20(3), 249-266. 37
12. Vasudevan, V., Kumaresan, V, and Nambisan, S.S. 2008. Development of a GIS-based 38 Safety Analysis System, Final Project Report submitted to the Regional Transportation 39 Commission of Southern Nevada, Transportation Research Center- University of Nevada, 40 Las Vegas. 41
13. Thomanna, Jose V. September 2006. Traffic Accident Data Analysis Using Geographic 42 Information Systems (GIS) Based Software. Institute of Transportation Engineers Journal, 43 76(9), 75-78. 44
Kumaresan, Vasudevan, Nambisan, and Duke
16
14. Clark County GIS Management Office (GISMO), 1 http://gisgate.co.clark.nv.us/gismo/gismo.htm, July 28, 2009 2
15. Model Minimum Uniform Crash Criteria (MMUCC), URL - http://www.mmucc.us/, July 28, 3 2009 4
16. Vasudevan, V., Pulugurtha, S.S., and Nambisan, S.S. 2007. Methods to Prioritize Pedestrian 5 High Crash Locations and a Statistical Analysis of Relationships between them. 6 Transportation Research Record 2002/2007, 39-54. 7
17. ESRI ArcGIS Version 9.2, http://www.esri.com/software/arcgis/, July 28, 2009 8 9
