mysociety / collideoscope-heatmap

Geoprocessing script for generating collision heatmap on Collideoscope.

collideoscope-heatmap

Script to generate the geopackage used for the Collideoscope heatmap (which originated here).

The geopackage is generated by combining road features from OS Open Road shapefiles with the points of all incidents in the Collideoscope database. Each feature in the output has a density attribute which represents the number of incidents per metre along the feature. Only road features with >1 incident (total) are included.

Quickstart

The simplest way to generate the geopackage is to use the Makefile to get you up and running quickly. Requires Docker and read-only access to the Collideoscope Postgres database.

$ export COLLIDEOSCOPE_DATABASE_URL="postgres://user:password@postgreshost:5432/collideoscope"
$ curl https://raw.githubusercontent.com/mysociety/collideoscope-heatmap/master/Makefile | make -f - heatmap

This takes an hour or so to run, and will generate heatmap.gpkg in the current directory.

Running via Docker

You gain more control over the behaviour of the script by invoking docker run manually. For example if you already have OS Open Roads downloaded locally to $HOME/Downloads and want to use that, and want the heatmap stored somewhere other than the current directory:

$ export COLLIDEOSCOPE_DATABASE_URL="postgres://user:password@postgreshost:5432/collideoscope"
$ docker run \
    -e COLLIDEOSCOPE_DATABASE_URL \
    -e OS_OPEN_ROADS_PATH=/input/oproad_essh_gb.zip \
    -e OUTPUT_PATH=/output/collideoscope.gpkg \
    -v /data/library/mapserver:/output \
    -v ${HOME}/Downloads:/input \
    -t davea/collideoscope-heatmap

Running locally

The script uses pipenv for virtualenv management, so install that if you haven't already. Then, run the script:

$ cat .env
COLLIDEOSCOPE_DATABASE_URL=postgres://user:password@postgreshost:5432/collideoscope
$ pipenv install
$ pipenv run python generate_heatmap.py