There was a gap in the market to recommend investment portfolios which driven by crowdsourced data! This inspired us to use personal financial data, and use it to help the community and learn from it about how to invest while still keeping it private!
What it does
In the end, it is a mobile app that acts as an aggregator service and provides a suite of features such as analytics. But our main feature is providing intelligent recommendations and ratings for your investment portfolio.
How we built it
We used mongo as our data store, flask as our backend server, and angular-native to build a mobile app. We also used numpy, scikit-learn to implement our models.
Challenges we ran into
We ran into the first challenge of getting the data from Yodlee. This proved very hard to mock in a "non-random" manner that simulates good real personas. We also had problems dealing with some charts on angular, but get eventually got it to work :)
Accomplishments that we're proud of
We got a really robust machine learning model, which uses a RL idea of crowd-sourcing data for getting better as we get more users. We also designed our model architecure to keep the recommendations personalized to you!
What we learned
Learned a bunch of new frameworks and libraries. Learnt a lot more about investing and various ways of doing so from the good folks at Capital One and Prudential