(Group is tentatively closed) This project will have an unsupervised learning portion and hopefully a supervised learning portion. The goal of the unsupervised learning portion will be to find "patterns" in sales of a large retailer's time series sale data (I have one in mind but many of these exist). The specific patterns we look for can be discussed between us but some immediate ones that come to mind are things like which time of year do products of certain types sell most often, which product sales seem to influence other product sales, which variables most influence a certain product's sale the most and many more which we can discuss. I expect a lot of dimension reductionality, clustering, and regression will be involved.

If the first part goes well, we can add a supervised learning component where we try to predict profit sales based under the assumptions of some of the patterns we've found.