Collect your data: go to ATM machines around town and dig through the trash for ATM receipts. Randomly gather 100 or so for each ATM, going out of your way to get a good geographical and bank distribution.

Now enter the balances, withdrawal amounts, and dates into a spreadsheet with the ATM’s location. Map the average account balances at each ATM’s location to get a picture of who’s got the money. Map the withdrawn amount to see who’s spending it. Map cash volume at different times of the day to see where people are spending at 8pm, at 4am, at noon.

Take daily samples from the same ATMs. Is there an ATM in the financial district that noticably follows the stock market? How about the one in the ‘hood that follows the MegaBucks lotto jackpot?

Take monthly samples from the same ATMs and adjust for inflation. Map who’s spending more or less (the rich or the poor) from month to month and year to year. Who spent the most last December, and who was holding back? Which part of town got hit the hardest by the last recession?