Help me make a map of response rate by zipcode

January 4, 2012 1:06 PM

What I want to do seems simple enough. But, it turns out, I may be too simple to do it! Or something. I have a dataset of survey responses. There are approximately 80K observations, which include response zipcode. What I want to do is make a map depicting which areas had the heaviest/lightest response rate. I have the following software at my disposal: GoogleEarth (but not pro) -- this did not seem to be up to the task; MapPoint (the very current 2004 version); and SAS

I've had some luck with SAS, in that I was able to create a map, but it didn't look great, or even, really, correct. This could, quite possibly, be user error. I've had less luck with MapPoint, but this seems to be the easiest way to do it, and I just don't know how. I'm able to create the simple pushpin map, but that's not particularly useful visualization.

There is some possibility of using ArcGIS, but my experience with that is pretty much non-existent, so that may not be useful.

Geographers and others-- your help is appreciated

posted by mcostanzo to Technology (9 answers total)

Tableau will make this pretty trivial. Watch the videos linked on their site, and you'll have this together in around an hour including the video watching time. As a bonus, you'll be able to quickly manipulate your map to visualize in different ways.
posted by bfranklin at 1:11 PM on January 4, 2012

Try using Many Eyes. You should be able to drop in a .csv or other spreadsheet of your data and create a map visualization of it. ArcGIS is the heavy-duty way to do it.
posted by The Michael The at 1:11 PM on January 4, 2012

BatchGeo looks pretty slick as well.
posted by lantius at 1:29 PM on January 4, 2012

Seconding BatchGeo. A coworker used it the other day to pretty quickly plot a bunch of zip codes on a map. It even aggregated them, such that if there were multiple of the same zip code, the dot on the map was bigger.
posted by losvedir at 1:33 PM on January 4, 2012

1. Create a table that has two columns: zipcode, and number of responses. Export as .csv, or .xls (not .xlsx)
2. Download the national zip code shapefile
3. Open ArcMap
4. Add zip code shapefile
5. Right-click on shapefile, click "Joins->Tabular join"
6. In the join dialogue, open your table from Step 1, select the appropriate zip code column from the shapefile table and your table. Click "OK"
7. Double-click on the shapefile, click on "symbology", create a classification based on quantity (left hand side of dialogue), based on the number of responses (available from pull-down menu). Edit colors as necessary. That will color your polygon according to how many responses you had. There's also an option in the symbology dialogue for creating dots with size relative to response #s. These instructions are for ArcView 8, 9 or 10 (i.e. not 3.x)
posted by one_bean at 1:37 PM on January 4, 2012

Google Visualizations can help with this.
posted by trunk muffins at 1:44 PM on January 4, 2012

You could do this super easily with Google Fusion tables as well- it will kick out a pretty heat map, among other things.
posted by rockindata at 2:30 PM on January 4, 2012

I've used Mapalist to convert spreadsheets in Googledocs to .kmz files that can be uploaded to Google Earth. I think it would work with just Zipcodes instead of full addresses.
posted by Potomac Avenue at 5:01 PM on January 4, 2012

Awesome. Thanks everyone. Lots of helpful tools to check out. Much appreciated.
posted by mcostanzo at 4:57 AM on January 5, 2012