Unpublished sources.

According to the "Unpublished Documents" chapter of Kyvig and Marty's Nearby History, (mentioned in my last post,) there are over 4 billion pieces of paper in the United States National Archives.
Any archives we'll be searching through won't have quite such an abundance, but it does put into perspective the almost nonsensical amount of documentation people (and especially the government) have left behind.
A look at the U.S. Census records is probably in order, since we're dealing with the displacement of a large number of people in the path of I-496. Demographic information of all sorts can be gained by wading through these records, however unenviable a task it may be.
I actually used the MSU archives once. I was writing a paper for a US history class and I was thinking of using some personal letters that showed up in an MSU library. What I found (after filling out a form and donning some white gloves) was almost illegible, totally unusable for a 10 page undergraduate research paper, and ultimately I was embarrassed at wasting the archivists' time in retrieving it for me. (For the record they were very nice and helpful.) I'll bet some people in our class (maybe even I) will be going in their in search of other records before March is over.