Geospatial Data Visualization Tools

[g4brielvs]

You have definitely seen a few data visualizations on Fourquare Studio. Opening this thread to discuss pros and cons of using it and other options for geospatial data visualizations.

Formerly known as Unfolded.ai Studio, Foursquare Studio was developed by the ex-Uber enginners behind Uber Movement (e.g., H3 and kepler.gl). AFAIK Uber Movement is now defunct. I have to confess I was not very thrilled with the acquisition by Foursquare, but Studio seems to remain a powerful hosted instance of kepler.gl with added funcionatilies (e.g., enhanced colormaps).

In particular, Studio enables map creation with H3 which results in hierarchical, uniform and performant visualizations. And maps can be embedded (e.g., Türkiye Earthquake Impact). That's not something exclusive to Studio, but the tool plays pretty well with it. Because the hexagon cells are about the same area, any H3-aggregated index promises be more comparable. And it is easy to go back and forth between the H3 tessellation and boundaries (I'm going to release a package that encapsules this nicely).

Now, I'm curious to learn from you and learn of other tools you use or have used at the World Bank for data visualizations. From QGIS, geopandas to the Geospatial Platform and beyond.

[andresfchamorro]

I think the main factors to consider are the type of input data (relates to the type of visualization) and the audience (end users).

From the examples you've shown, it looks like foursquare is particularly good at handling timeseries gridded data. The maps looked great and performed well. And it's only for internal demonstration, so that makes sense. The main con is that we can't use it on WB websites (not sure what it would take to get it approved).

Here are some other tools/plaftorms, definetely nonexhaustive. Hope it's useful to have them all listed.

WB Platforms

• Geospatial platform: Managed by Geolab/ITS, pre-loaded global layers, teams can look for satellite imagery
• GEOWB: ArcGIS online maps, easy to set up, webmap can be shared internally
• Geospatial Poverty Platfrom: New site from PIP team for harmonized poverty data

Mapping Tools

• QGIS/Arcgis Pro/ArcMap: Software solutions
• geopandas/matplotlib/contextily: Enough for static maps in python
• folium/leaflet: Interactive maps quickly from python notebooks
• geemap: Great if you work with GEE and python
• R Shiny: I believe this is a solution supported by IT and teams that use R (DIME, Poverty)
• Mapbox: A lot of external companies have used this (i.e SDG Atlas, Covid dashboard)
• D3.js/leaflet.js: Hard to set up, very specialized skillset required

[ivannovazzi]

Worth adding Moveet to the list if your use case involves visualizing moving vehicle data rather than static geospatial datasets.

It's an open-source real-time fleet simulator: it loads a road network from OpenStreetMap/GeoJSON, runs A* pathfinding to generate routes, and streams live vehicle positions over WebSocket. The built-in dashboard uses D3.js SVG rendering (no Mapbox/Leaflet) to visualize the fleet in motion — heat zones, routes, and POIs included.

Useful as a realistic data source for prototyping mobility dashboards or testing how tools like kepler.gl handle streaming location updates from a fleet.