We want to make a graph with a point for every .edu . The size of each point will be proportional to the number of academic papers on subject TOPIC.

Here's what you do.

1. Get a list of all .edu's and their long/lat coordinates. The easiest way to do this is to use something like the ip2location.com database or maxmind.com. You'll find there are about 2,000 different .edu's.

2. We now have a list of all .edu's and their location on a map, now to determine the size of the points. Use the Google API to run a google scholar query of for every .edu:

   http://scholar.google.com/scholar?q=site%3ASOMEEDU.edu+TOPIC

   If you were interested in the generic distribution of academic PDFs without regard to a topic, just leave the TOPIC field blank, for example:

   http://scholar.google.com/scholar?q=site%3ASOMEEDU.edu

3. Count the number of hits you get, this will determine the size of the data point.

4. Plot using Google Maps. The geo coordinates determine the location of each point, the number of hits on Google Scholar determines the size/color of the point. The Google Maps API supports doing both out of the box.

5. Done.