Recently I joined a web seminar. It was about using geospatial files in Tableau – the new feature released in Tableau 10.2. We were not able to use geospatial files directly in Tableau before, but only through some clever workarounds. So I made it my goal to learn how to create these files and visualize geospatial data in Tableau itself.
Now that I’ve mastered this skill, I want to share my experience with you.
Where to get geospatial data
There are many ways to make geospatial shape files for Tableau. For starters, you can export them from your Google Maps or use the GMap GIS tool for very simple shapes.
Tableau allows you to import geo data in the following formats:
KML files (.kml)
ERSI Shapefiles (.shp)
MapInfo Tables (.tab)
MapInfo Interchange Formats (.mif)
GeoJSON files (.geojson)
You can obtain publicly available geospatial files from the following sources recommended by Tableau:
From the US:
Data.gov: full catalog of open data in the US
Census.gov: census boundaries
Open Data Philly: great example of what many major cities are offering
National data portals:
UK Ordnance Survey: local administrative and electoral boundaries for Great Britain
Swiss Open Data: administrative, conservation, utilities, and more spatial data for Switzerland
Instituto Brasileiro de Geografia e Estatística (IBGE): comprehensive administrative and census data for Brazil
GADM: catalog of administrative boundaries for nearly every country in the world
HDX (Humanitarian Data Exchange): makes humanitarian data easy to find and use
Creating my own geospatial files for Tableau (Geospatial IronViz)
If you are anything like me, you’ll want to make your own complex files. To do just that, I downloaded ArcGIS (Windows) and QGIS (Mac) and gave both a fair trial. I used each for 30 minutes, making some test geo shapes. Eventually, I found QGIS to be my best candidate and went forward with making my geo shapes using this open source software.
I've always been interested in cartography. Recently I helped to digitize the David Rumsey map collection by taking scanned map files of old maps and mapping them onto the modern world map using coordinates (where available) or by matching features such as lakes, rivers and cities.
So I decided that for my geospatial training I would take a few really early maps, recreate them in QGIS, and visualize them in Tableau.
I selected these maps and web resources:
Each map is an improvement on the previous map, so instead of recreating each map, I highlighted only the improvements made by cartographers, because that's what matters. There are many maps out there, but I only needed those that actually provide better information.
It’s amazing how in the past people would bet their lives on a piece of fabric, paper or skin with a few very simplistic graphics shapes on it. Those adventurers shared their findings and inspired one another to continue the discovery. It reminds me of the data discovery process that we all, as data people, go through. So I used this analogy to create this minimalist data story.