Map Roulette Geojson Generation

To generate the geojson, you'll need to detect the cases, gather the required data, and define how the geojson should be created.


You will have to design an algorithm that can detect the cases that you are looking for. Once you are able to detect the cases, you now need to capture the useful data.

For an example of existing detection please see the function bool IsUnroutableNode in

Gathering the Data

At a minimum, you will need some sort of geojson feature to display on the Map Roulette interface. This can be a point or linestring or a combination of them. At some point in the data capture you must capture a lat-long pair (PointLL) that is unique for each case you detect. This will be used later as the identifier in the geojson. You will need to modify the roulettedata struct to hold your new data.


To have your geojson correctly recognized by Map Roulette and the automation scripts you will need to have a specific format. You will need to implement something like this that creates your geojson and adds it to a collection. A single task's geojson should look something like the following:

    "instruction": "This node is either unreachable or unleavable. Edit the surrounding roads so that the node can be accessed properly",
    "properties": {
        "key": 4770031602477192848,
        "type": "Node"
    "type": "Feature",
    "geometry": {
        "type": "Point",
        "coordinates": [8.64613, 44.63913]

The properties.key field here is the hashed PointLL (lat-long) used as the id. Look at the previous link about geojson creation to see how to create the hash.

That should be it, now you can start automating your Map Roulette tasks.