Did you know that in SPSS Statistics you can make maps? And it is actually fairly straightforward! Back in the nineties SPSS (version 4!) had some mapping capability thanks to a collaboration with MapInfo, then they dropped the mapping, but since SPSS 20 mapping is back. Here we show how to map your SPSS data.
SPSS Statistics comes with a few map templates – basically global country boundaries and US states. For the UK you can get your own administrative, census and other boundaries from the UK Data Service Census Support as Shapefiles. Make sure your Shapefiles has a unique identifier for each area, like a county name or a census ward code. Then import the Shapefile into SPSS Statistics (Utilities > Map Conversion Utility). See the very good SPSS help for details.
To make a choropleth map in SPSS Statistics for a scale (numeric) variable:
- Make sure your SPSS table also has a variable that identifies locations, – e.g county name or census ward code. Let’s call it LocationID.
- Go to Graphs > Graphboard Template Chooser > tab Detailed > Choropleth of Sums.
- Set: Data Key = LocationID, Color=<AnyNumericVariable>, Data Label = LocationID
- Click Select a Map File. For Map select your map template, for Map Key select the equivalent of your LocationID. Click Compare. All keys should match, meaning that for each LocationID SPSS finds a corresponding location in the map.
- Click OK and OK again. The map is created in the output viewer.
- To change colors or styling, douple-click the map, it opens in the Graphboard Editor. Click anywhere in the map to select the polygons. At the bottom of the window go to the Color tab and select the colors that you want for low and high values.
As you probably noticed there are lots of map types, like bar, line and pie charts, or ‘Coordinates on a Reference Map’ (Latitude/Longitude from variables!)
I found a few issues, for example, ‘Choropleth of Values’ only works with categorical variables, not numerical ones (why?) and Bar, Line and Pie chart maps can use only a single variable (what??). Still, if you have your data in SPSS anyway this is a great way of making maps.