Osmlint to detect errors in OSM

Posted by Rub21 on 2 March 2016 in English (English)

We have been developing OSMLint as a tool identify common geometry and metadata problems OpenStreetMap data. OSMLint processes vector tiles with turf.js and TileReduce to produce geojson files of detected data errors.

In this post, I will demonstrate using OSMlint with OSM-qa-tiles to detect errors in OpenStreetMap.

  • Clone my sample repository and install required software.

git clone

cd run-osmlint


  • To run this for your country, run the The requires 2 parameters:

sh ./ "[-73.975067, -34.108073, -34.512177,5.5056583]"

The script runs all the available linters and produces a gist file that you can visualize.

screen shot 2016-03-01 at 16 41 19

If you want to run a specific linter, check out sample commandes here:

We are continuing this work by adding more linters, if you have any suggestions, check out the code or open a ticket.

Comment from d1g on 4 March 2016 at 23:45

Lovely dashboard! "Filter data" is a gem, but unexplained/no docs.

I would recommend to explain how to write and process amenity=cafe+name=Cafe Mirage -> amenity=cafe+Mirage using OSMLint - so users are not afraid of new interfaces and workflows (because common tasks are the same).

Could you please create sample code how to detect disconnected graphs (connected components with less 20 ways)? It is possible?

More platform independent code would be nice in examples (Java / Python)

Comment from Rub21 on 9 March 2016 at 15:51

Hello d1g:

There are basic examples of code:

if you want to filter data, you can only use if sentences

About your question amenity=cafe+name=Cafe Mirage

here is the example:

The result on gist :

The result on dashboard :



Comment from Rub21 on 9 March 2016 at 16:02

Hey d1g, For the above the example, I used peru.mbtiles

osmlint namecafeonamenity --bbox="[-81.462250,-18.417079,-68.542328,0.24169850]" peru.mbtiles>peru.json

Comment from d1g on 8 April 2016 at 18:56

Yes, thank you for examples in repo!

It is easier to learn new tools when you can compare apples from one tool to apples in other tool (and especially if it was done before).

Login to leave a comment