NessieVIS

Inspiration

We saw a cool visualization of Venmo paper-trail data and decided to use Capital One's Nessie API to simulate customer behavior and purchasing decisions with hierarchical Bayesian modeling.

What it does

Simulates and generates data about customers and which types of products they purchase. Analyzes data and uses LDA clustering to rediscover the underlying connections between different merchants.

How we built it

We used Capital One's Nessie API in Python to write a variety of functions for adding and measuring purchases between customers and merchants. Using the random module in num.py, we generated purchase data. We then used the gensim package to fit the data. Finally, we used d3.js in Javascript to create visualizations of the purchase data.

Challenges we ran into

We had no data, so we had to figure out a reasonable model for simulating interesting and realistic data. Improving runtime when processing large amounts of data was also a challenge.

Accomplishments that we're proud of

Developing functions for generating random names of people and of stores Creating realistic results that measure the connectedness between merchants

What we learned

We gained a better understanding of scripting in Python and writing modules between a variety of different data types. We also learned a lot about interactive data visualization in Javascript.