We collected web archives of ads every Sunday for a one year period from October 2012 to September 2013. The number of ads grew over the study period with approximately 50 ads on a typical Sunday in October 2012 to 240 ads on a typical Sunday in September 2013. We found we needed to do weekly archives because the older ads would disappear after about 2 months. Fortunately, because the ads were organized into a standardized format for listing information about the unit, we could employ a webharvesting application to automate the conversion of textual information in the ads into a spreadsheet database. With the original raw collections of 7312 ads, we then cleaned up the data deleting ads that were duplicates, that were not housing but storage spaces, and those that were missing key data needed in our model such as the unit's size, price, location, etc. In the end, we culled 3385 ads, retaining about 46% of the initial harvest.