Inspiration

We found that businesses needed a tool to analyze and compare discounts and sales. Square's API provided us with the tools to do so.

What it does

When you log in with Square, you can compare your discount revenue to your baseline revenue graphically.

How we built it

We used Node.js on the back end, and JavaScript/HTML/CSS on the front end. We take samples the collection of sales for each discount and the baseline several times to gather a normal distribution for each sale (see: Central Limit Theorem). We then allow you to compare these distributions and see what works for you.

Challenges we ran into

Square doesn't allow you to create retroactive data, so we had to make a CSV file of data similar to the format Square would give us instead of using data from just this weekend. In our current implementation, the code that interacts with Square's API is commented out, and we're demoing using data that we generated. Additionally, we were hoping to create a predictive model of the discounts' effect on price using a linear regression, but we found that the data wouldn't be suitable for the technique (multicollinearity, not a correct distribution). We still found it useful to see what worked in the past and make recommendations on that data.

What's next for DealData

We'd like to finalize bug fixes and integrate the visualization properly. We'd like to add a few more functionalities, including a regression model and more specific comparisons (e.g. certain discount on a Friday vs for the week). We'd also like to test directly with the API using a real business's data.

Share this project:

Updates