This project makes traffic information integration into GraphHopper possible. In the example we are using real time traffic from Cologne and update every ~2 minutes, the UI will update too. The web UI is basically a slightly changed GraphHopper Maps UI with the traffic data in a separate HTML5 canvas layer:
The nice thing is that the routing changes immediately after recieving the traffic data, i.e. in real time.
Note: When using the speed mode (prepare.chWeighting=fastest) real time is not possible for large areas as you'll have to prepare the data again after new data arrived which will take roughly 9 minutes for Germany. So only near-real-time.
Three simple steps:
wget http://download.geofabrik.de/europe/germany/nordrhein-westfalen/koeln-regbez-latest.osm.pbf./td.sh datasource=koeln-regbez-latest.osm.pbf- visit http://localhost:8989 to try routing in the UI and see traffic infos
There is an endpoint to fetch all roads with changes which is used for the traffic info:
Disable the update for Cologne in the source, pick your location and start the server for your area:
./td.sh datasource=your-osm-file.pbf- visit http://localhost:8989 to try routing in our UI
Now feed some data to '/datafeed':
curl -H "Content-Type: application/json" --data @traffic.json http://localhost:8989/datafeed... and try routing again. Note, in order to use the provided example traffic.json you'll have to use the specific area, get it here
The data format is very generic and can be used for other information influencing routing:
[{
"id": "1",
"points": [[6.827273, 51.190264]],
"value": 10,
"value_type": "speed",
"mode": "REPLACE"
}, {
"id": "somethingelse",
"points": ...
}
]Note, the point list is in geo json and therefor use lon,lat instead of the more common lat,lon order
This code stands under the Apache License 2.0
The user interface can be modified via npm, then do:
npm install
npm run watchThe traffic information is rendered in map.js via fetching all roads from the /roads endpoint