Mpire Buyer is a shopping service that provides comparisons between the prices of new and used items. I developed a lot of the Buyer UI as well as a distributed system named Marissa that can quickly compute analytics on historical auction data that Mpire Buyer uses this to provide historical pricing information.

We also use Marissa to implement an algorithm that uses historical auction data to find the most relevant items for a given search query. This is very important when you're trying to, for instance, say something about the average price of a good on eBay. In order to do so, you need to be sure that you're comparing apples to apples. If somebody searches for iPod Nano, you can't include iPod Nano armbands in the average price statistic or else it will pull down the average and make it useless. Try it out and see for yourself: I'll think you'll find that our relevance algorithm to be uncannily accurate. Try opening up an eBay window side by side and compare the quality of our results.

To the right you'll find a link to a screencast of Mpire Buyer that shows off some of the major features of the application.