RiskCanvas

Inspiration

We were inspired by stock predicting models and online financial advisers. We wanted to enhance this with AI.

What it does

Product uses a PPO algorithm to analyze user financial risk desires and current market trends to give recommendations on stocks in invest in.

How we built it

We built it using Python.

Challenges we ran into

Some challenges included properly training our AI algorithm to look at current market trends and give valid output recommendations. The model at first had a bias towards moving away from the desired user risk, so we had to calibrate this.

Accomplishments that we're proud of

We are proud that the model finally ended up working and that our UI has a clean final look.

What we learned

We learned how to use APIs such as polygon to provide data to create recommendations for users and efficiently train RL models.

What's next for RiskCanvas

To continue our project, we will train our model with larger data sets over longer time periods to give more accurate suggestions based on the market. We will also include capabilities for futures, commodities, forex, and crypto. Also, we want to give users the ability to directly connect their investment account(s) to our program, so that they don't have to manually input all their positions.