crashstats's Introduction

Gather Crash Statistics

Data from https://unfallatlas.statistikportal.de/ converted the SHP to GeoJSON with QGIS. (Don't like to add geotools as a dependency ;))

Results

Using OSM data from early 2019 and crash data from 2017, the total length is approx. 13000km or if you count the motorways as "two oneways" like we do in the code it is approx. 26000km
In Germany approx. 17000km of 26000km (total length) of the highway network are without a speed limit, i.e. there are at least 65.4% of the motorways without a speed limit
On road segments of a motorway without a limit there are 69.5% of the people die, 65.4% are badly injured and 57% are lightly injure.

TODOs

When calculating the percentage of dead or badly injured people we need to consider traffic count data
use older openstreetmap data to better match the situation for the older crash statistics
traffic sign data in openstreetmap might be wrong, check this by a view examples

Installation

The following steps can be omitted once 0.12 is out:

git clone https://github.com/graphhopper/graphhopper/
cd graphhopper
git checkout 843f994
mvn clean install -DskipTests=true

Now create the crashstats project and run it:

git clone https://github.com/karussell/crashstats
cd crashstats
wget https://download.geofabrik.de/europe/germany-latest.osm.pbf
mvn clean install
java -Xmx3g -Xms3g -Dgraphhopper.graph.location=graph-cache -Dgraphhopper.datareader.file=germany-latest.osm.pbf -jar target/crashstats*.jar
# open browser at localhost:8989/maps/

Recommend Projects