We all forget things. Some of us forget toilet paper. Others forget to grab that loaf of bread on the way home from work. Wouldn't it be nice to have someone take care of that for you and enjoy time with your family?
What it does
The software tracks purchases made from a customer's Capital One account to see how often certain items are bought. Using this purchase history, Bag to the Future can automatically remind the customer what to buy the next time they're out shopping. We wanted to accurately model the purchase of items, so we used an exponential distribution for each product to predict the probability of purchase on any given day. Then, we calculate the average time span between purchases of each item and comparing it to how much time has elapsed since the last purchase. The website given below simply models how the reminder system would work over a year. The current day of the year can be seen on the console.
How we built it
Due to the fact, that Capital One's database was lacking sufficient data to track purchase history, we made our own model of customer transactions and pushed it to Capital One's database. This data was then processed as described above to find trends and then compared against current situations.
Challenges we ran into
The mathematical model to predict whether a customer purchases a given item took the bulk amount time. We eventually went with a piecewise exponential distribution to model the probability a customer makes a purchase.
Accomplishments that we're proud of
We accomplished our goal to automatically set reminders to buy groceries and other items as households run out periodically.
What we learned
We learned a great deal about Capital One's API. This was our first time processing so much financial data. We also learned about consumer behavior in changing markets.
What's next for Bag to the Future
We would like test our application on real data to see how effective our product will be in the complex consumer market. We could also collaborate with organization such as Groupon and large scale retailers to perhaps provide unique coupons and incentives to help customers save money. We could also look into purchase automation and buy certain items after the average interval length has passed.