Transcript CARE - Past Traffic Records Forums Index

GIS Display and Analysis of Crash
Data
Gautam Mistry
University of Alabama, Tuscaloosa
Prepared for:
29th International Traffic Records
Forum
July 15, 2003
Objectives
– Mapping crashes stored in Critical Analysis
Reporting Environment (CARE) on maps using
GIS
– Spatial Analysis of mapped crashes to identify
patterns and hotspots
CARE
Critical Analysis Reporting Environment
– Database maintained by the state of Alabama
– Detailed information about vehicular crashes
– Information can be retrieved by “data mining”
– Crashes presented in tabular form and graphically
in a linear form
CARE

Crashes can be selected meeting specific criteria

Crash locations verbally described without coordinates


State routes are graphically represented by a line and
number of crashes represented as stack of points at
mileposts shown as numbers
Graphical display does not correlate with roadway
features such as curves, bridges, rail-road crossings,
and etc.
CARE Location Data

CARE stores crash locations in the form of nodes
(intersection), links (roads), and route milepost
(highways)
– Crashes at intersections are identified by node
numbers
– Crashes on roads are identified by three fields: Link_ID,
Node_1, Node_2
– Crashes on highways are identified by the route and
milepost at which the crash occurred
Creating Common Fields
To connect the CARE data to the GIS roads layer, a
common field is required to be created for nodes
and links
– Value of Node_1 is inserted in the column Connect_ID to
represent crashes on intersections
– The three fields describing links are concatenated to a
single entity and placed under a separate column
Connect_ID to represent crashes on links
– A table containing crashes were directly added to the GIS
file as an event table to represent the crashes in mileposted highways
GIS Map

Nodes layer was created, and node numbers
were assigned for each intersection

Link number (Link_ID) and node numbers
(Node_1, Node_2) were manually inserted in the
Roads layer for each road segment

Routes were created for highways and milepost
were assigned to the ends of the routes
CARE Linked to GIS

Node_ID column from nodes layer was related
to Connect_ID column from CARE

Connect_ID column in CARE database was
related to Connect_ID column from roads
layer in GIS for crashes that occurred on links

Highway crashes are mapped on the routes
created in GIS by dynamic segmentation
BASE MAP
Tuscaloosa County
®
Spatial Analysis

Hotspots are identified to indicate the
locations with high crash frequencies
using thematic mapping and buffering
Thematic Map for Links
®
Legend
Number of Crashes on Links
0
1
2
3
®
Thematic Map for Nodes
")
+
$
Legend
Crashes at Nodes
1-5
6 - 10
")
11 - 15
+
$
16 - 20
®
Crashes on Links within 0.5-mile Buffer from
Major Highways
Legend
0.5-mile Buffer of Major Highways
roads
Number of Crashes
0
1
2
3
Spatial Analysis

Crashes mapped on GIS can be spatially
correlated to roadway features such as
bridges, railroad crossings, crossroads,
and etc.
Thematic Map of Crashes at Rail-road Crossings
!(
!(
+
$
!(
#
*
®
Legend
!(
roads
rail
Crashes at Railroad Crossings
!(
!(
#
*
")
+
$
1
2
3
7
Crashes on Links near Rail-road Crossings
Legend
®
roads
rail
Number of Crashes on Links
0
1
4
Spatial Analysis

Crashes on Interstate and State
Highways were mapped according to
the milepost at which the crashes
occurred using dynamic segmentation
Crashes on I-20/59 According to Milepost
Tuscaloosa County
!(
!
!(!((
!
!( !(!((
!(
!( !( !(
!( !( !(!(!(
!(
!(
!(
!( !(
(!!(
!(
(!!(
!(!(
!(
!(
!(
!(
!(
!( !(
(! !(
!( (!
(!
!(
!(
!( !(
!( !( !(
®
!(
(! !(
Legend
!(
!(
(!!(
Crash Locations
I - 20/59
county
Conclusions




This GIS application allows user to view the crashes
on road map, which gives spatial representation of
crashes
It enables CARE data to be correlated with existing
roadway features like bridges, rail-road crossings,
and etc.
It enables identification of hotspots and crash
patterns at nodes, links, or route mileposts
Use of multi-year data helps in graphical
representation of before and after analysis
Questions