Let's Map Some Points of Interest in San Francisco City using Mapillary!

Posted by poornibadrinath on 25 January 2017 in English (English)


When you happen to walk on a street, what usually catches your eye? The street? The trees? Or a cute little shop in the corner having a neon sign board saying pastry? Shop with a pastry sign, mostly! Such shops, stores, restaurants, bakeries, ATMs, etc. are what we call Points of Interest (POI). POIs makes the map well, interesting (pun unintended). It is equally important as streets and buildings to any map for navigation.

Mapping POIs usually requires a mapper to collect GPS traces and take photos on the ground. With Mapillary street-level photos, anyone familiar with the city can look at any Mapillary photo taken by others and use them for mapping POIs.

Here's how you can map them👇🏻

  • Open an area you are interested in JOSM.
  • Enable the Mapillary plugin to see the images of that street. If you don't have the plugin available go to Preferences > Plugin.


  • Use the Mapillary filter to get only the latest photos. This prevents us from looking at old photos that shows POIs that may not exist on the ground anymore.

sfmapillary Enabling a Mapillary filter will help in prioritising only recent images.

  • Check every street for POIs. Found anything missing? Add them in OpenStreetMap. If they are already present on the map, verify or add more information.

  • To make POI's appear more prominently, use the osmic map style. osmic

Mapbox has added millions of photos in Mapillary last year and continues to add more!

Our streetcam rig.

The coverage in San Francisco City is nearly complete and we are starting to improve POI coverage using these photos.

screen shot 2017-01-25 at 12 30 23

Head over to our ticket for detailed workflow or jump right in and grab a task.

Mapping POIs is straightforward, but maintenance is challenging. Even though we are limiting to only recent Mapillary photos, a lot can be changed within a span of days. In such cases, ground truth matters and local knowledge would be the best to know where what has been changed.

If you have any feedback on the workflow or suggestions on improving the same, please do let us know. Looking forward to the continued collaboration.

Thank you,


From Mapbox Data Team

Login to leave a comment