gCensus

I recently completed my Masters of Science in Computer Science at The University of Mississippi. One of the classes I took during my last semester was a course concerned with processing large quantities of data. Specifically, the course focused on scientific data collected in such fields as fluid dynamics, physics, and weather. The class was small, about a dozen people, and this allowed the class to be somewhat informal in nature. We read two or three published research papers per week and discussed the papers in round table discussions during class meetings.

One of the assigments we were given was open-ended. The assigment was to "do something interesting with a large data set". I looked around at possible datasets and came across the census data. At approximately ten gigabytes, I thought it qualified as a large data. At the same time, spring 2005, Google had just released its map technology. I thought the two were a perfect match and gCensus is the result.

No, gCensus is in no way associated with Google. Google has not given its blessing nor has Google been consulted about this project. Google is in no way responsible for any of the content on this site.

There is a very large amount of information available at some zoom levels for some regions.  If I attempted to load all of this data, many computers and browsers would run very slowly.  Therefore, I made the decision to show a maximum of 50 data points for any single map screen.  I also made the assumption that the data a user would be most interested in would be in the center of the map. Therefore, all of the data points are clustered into center of the map. Even though all data points are not shown in the map, the totals located to the right of the map are complete for the viewable area.

The US census bureau recognizes several geographic areas for their data.  These areas scale in size from the entire United States down to the block level.  For this data I use the following 7 geographic data levels:

1. Entire United States
2. Regions
3. Divisions
4. States
5. Counties
6. Census Tracts
7. Block Groups
8. Census Blocks

Blocks are the smallest unit of land recognized by the bureau and can conceptually be thought of as city blocks.  However, in rural areas census blocks can be several square miles in size and contain no people.  Block groups are made up of several blocks, tracts are made up of several block groups, counties are made up of tracts, and so forth.

Google maps using Javascript, XML, XSLT, and CSS to create a web application that behaves more like a desktop application. With normal web programming, the user is presented with a form, the user inputs data, the user submits the form, and finally the entire web page reloads with new content. AJAX, as this type of programming has been coined,  does not require the reloading of the entire interface.  The Javascript is able to make calls back to the server to get the required information and update only those parts of the interface that needs to change.

In the case of gCensus, every time the map is panned or zoomed the Javascript contacts the gCensus server and requests information for the viewable map region.  At the same time, the Google code contacts the Google servers and requests the required images needed to display the map.  The page is only reloaded when the user clicks the refresh button.