3 weeks ago
Is it possible to display a geo map in the style of a heat map using SAS VA? I am currently using a bubble map and want to show the density of of a population by zip code.
Alternatively, is there a way to bring the largest bubbles to the front of the map? When I look at things by zip code the largest bubbles tend to be overtaken by smaller bubbles making it very difficult to interpret.
Any suggestions or best practices/papers are welcome for SAS VA Geo Maps
3 weeks ago - last edited 3 weeks ago
You can create a region map where the regions are shaded by frequency. I think this is what you mean by a heat map.
However, you cannot do this for ZIP codes out of the box. There is a workaround to add ZIP codes (or any other geographic regions) to VA, but it is supported for VA 7.4 only. In theory it works for other VA releases but it is supported only for 7.4. It is also a fairly complicated administrator task.
See the documentation here:
About the bubbles, I don't think there is a way to bring the large bubbles forward. If the number of bubbles is an issue you might want to create a filter or rank. Zooming the map might also help if that is a possibility for you.
In general, mapping ZIP codes can be problematic because of how small they are in urban areas and how they can be very irregular and noncontiguous.
I hope you will find this helpful,
ETA: I am actually not 100% sure you can add ZIP codes in this way. You can add US counties and postal codes from other countries, but US ZIP codes might be a special case.
3 weeks ago
While we don't support heat map / density plots just yet (planned for future releases) - I found it best to use the geo coordinates plot for high cardinilaty maps. So things like US zip codes with over 40k values typically render better with smaller data points. The following screenshot shows an example for US with 80% transparency:
Zooming into the east coast reveals more details:
As Sam mentioned - you may also consider using regional overlay if you have access to the boundary data. Given the number of zip codes however you may not be able to render all data points for performance reasons. The following examples shows Florida:
Hope this helps. Regards, Falko