Regional Crash Analysis Utilizing GIS to Improve Regional Crash Analyses And Data Visualization

1
Regional Crash Analysis Utilizing GIS to
Improve Regional Crash Analyses And Data
Visualization

• Melissa Baldwin
• Transportation Planner
• Area Plan Commission of Tippecanoe County
• Lafayette, IN

2
Tippecanoe County

• 504mi2
• Population
• 163,364
• 9.8 in GQ
• Housing
• 65 Single Family
• 35 Multi Family
• Top 3 Employers
• Purdue University
• Subaru /Toyota
• Wabash National Trailers
• MPO Staff
• Executive Director
• Transportation Division Director
• Senior Transportation Planner
• Transportation Planner
• GIS Tech (vacant)

3
Purdue University

• West Lafayette Campus
• 39,000 Students
• 9 Part-time
• 12 International
• Largest University housing operation in the
  Country (31 of student body)
• Major Highways
• US231 East campus edge
• SR26 Divides North and South Campus

4
Road Network

• State Roads
• I-65
• US52, US231
• SR25, SR26, SR28, SR43
• 84 of Principal Arterials
• Urbanized Area
• 78.5mi2 (15.5 of County Total)
• 41 of road miles
• Safety Issues
• Driver inattention
• Wide and offset intersections
• Lack of access control
• Pedestrians
• Rural County Roads
• 74 paved and 26 gravel
• Safety Issues
• Deer
• Horizontal curves (90 jogs around fields)
• Vertical curves and sight distance

5
Multimodal Transportation

• Transit (CityBus)
• 2nd Largest ridership in IN
• 4.7M Annual ridership
• Rail
• Major Lines
• CSX
• Norfolk Southern
• Amtrak
• Numerous industrial spurs and yards
• Railroad relocation in 1990s
• Pedestrian/Bike Facilities
• On-road facilities
• Recreational trails

Then
Now
6
Indiana ARIES (Automated Reporting Information
Exchange System)

• State Crash Record Database
• 100 digital in CY 2003
• Electronic reporting
• QC Checks
• Linearly referenced to INDOTs road network
• MPO Access
• e-Report review
• Data extracts
• Report generation

7
Tippecanoe Co Annual Crash Report

• Started in 1989
• Report Contents
• Countywide Crash Summary
• Intersections for 10 Crashes (100ft rule)
• Crash Frequency
• Crash rate per MEV
• Critical Rate Factor Comparison
• Pedestrian
• Bicycle
• Motorcycle
• New GIS analyses Spatial Patterns
• Density Analyses (ArcGIS Spatial Analyst)
• Hot Spot/Zone Analyses (ArcGIS core and ArcGIS
  Spatial Analyst)
• Crash Density Visualizations (ArcGIS Spatial
  Analyst and 3D Analyst Extensions)

8
Tippecanoe County Crash Statistics

• Yearly Statistics (approximate)
• 1,700 Private property crashes
• 5,700 Roadway crashes
• 18 Injury crashes
• 1.3 Fatality and incapacitating injury
• 14-22 Fatality crashes per year
• 56 Tippecanoe crashes correctly geo-referenced
  by Indiana ARIES
• Local Processing
• Assign intersection status
• 100ft
• 250ft (used by Indiana 5 report)
• Mid-block
• GIS Scripts Intersection Statistics
• Assign crashes to intersection
• Compute statistics and rates

9
Density Analysis

• Density Summaries Useful for
• Overlapping crash points
• Data privacy concerns
• Hiding geocoding errors
• Smoothed data for 3D visualizations

10
Density Analysis

• Drawbacks
• More data, since continuous raster
• Must determine suitable grid spacing and search
  radius
• More work to display data and legend
• Different patterns for different search radius
  and coloring scheme

SR25 N, Crashes per 0.1 mile (CY 2002-2007)
Hot Spot
2D Top View
3D View Looking SW
11
Spatial Pattern Analysis

• Is the crash data
• Clustered
• Dispersed or
• Randomly distributed
• Is there significant spatial clustering of
• Severe crashes
• Deer-related crashes
• Run-off-the-road crashes

12
Hotspot Analysis

• Mapping Clusters -gt Hot Spot Analysis
• Identifies statically significant clusters of
  points i.e., those with values higher in
  magnitude than you might expect to find by random
  chance
• Hot Spot tool determines spatial hot and cold
  spots
• Computes a z-score (standard deviation)
• For 95 confidence level the z-score threshold is
  1.96
• -1.96 z 1.96 statistically no pattern the
  expected pattern is one of hypothetical random
  chance
• -1.96gtzgt1.96 statistically too unusual to be
  just another version of random chance

13
Crash Severity Hot Spots

• 18 Injury crashes
• 1.3 Fatality incapacitating injury of which
• 6.4 Pedestrians
• 4.4 Bicyclists
• 21.2 Motorcycles
• 58 at Intersections

(Data Ranges from 1-29 Crashes)
14
Fatality and Incapacitating Injury Crash Hot Spots

• Conclusions
• 45 of fatality crashes are occurring in
  statistically significant clusters
• Need to incorporate additional years worth of
  data

15
2006 Deer Related Crash Hot Zones
16
Run-Off Crashes 2006-2007

• Likely circumstances
• Horizontal Curves
• Vertical Grade
• Road Type
• 2 occurred on gravel roads
• Surface Conditions
• 52 Dry
• 33 Wet
• 14 Ice

• Unexpected Hot Spot
• Slight horizontal curve
• Fairly new pavement
• 55 mph
• Possible safety Concerns.
• 50 occurred during Wet surface conditions or
  Speed to Fast for Weather

Symbols Crashes Per 1000ftx1000ft Shading Hot
Spots (zgt1.96)
17
Run-Off Crashes 2006-2007Impaired/Speed/Failure
to Yield
Rural hot spots are centered around horizontal
curves
50 Speed and 50 Speed too Fast for Weather
100 Speed too Fast for Weather (Rain)
18
Crash Visualizations

• ArcGIS 3D Analyst Extension
• Crash Density Perspective Data Views
• Comparison with Land Use/Land Cover
• Comparison with Topography
• 3D fly through animations
• Flight Paths
• Captured animations into a media file

19
Conclusions

• Indiana MPOs can quickly and easily incorporate
  GIS-based safety analysis in the planning
  decisions
• Hot Spot Analysis can help you identify spatially
  significant areas for further investigation
• The public loves the density mapping and
  visualizations
• Future work
• More traffic enforcement related analyses (e.g.,
  right angles crashes at signals)
• Land use vs. deer crash analysis with local
  wildlife biologists
• Normalize crash densities with VMT from traffic
  count GIS layers
• DEMO 2006 Crash Animations
• - QUESTIONS and COMMENTS -

20
Density Analysis ArcGIS Tools

• Point Statistics
• Consolidate point data to a non-continuous
  gridded area
• Kernel or Point Density
• Consolidate point data based on density to a
  continuous gridded area
• Steps
• Geolocate reports on your network
• Decide a grid spacing
• Decide on the units for your density calculations
• Set the Radius of Influence 1 (no smoothing of
  data to adjacent cells)

21
Additional Information

• Several Analysis tools available
• ArcGIS and Spatial Analyst Extension
• CrimeStat
• GeoDat
• Good References
• Chapter 4 Data and Information for Considering
  Safety in the Transportation Planning Process
  http//tmip.fhwa.dot.gov/clearinghouse/docs/safety
  /chapter4.stm
• Mapping Crime Understanding Hotspots
  http//www.ojp.usdoj.gov.nij