goldmanHachs

Inspiration

Making investment decisions can be overwhelming and time-consuming given the large variety of financial metrics to pull from. What the average investor really wants is a simple, informed recommendation of stock purchases.

What it does

Given a ticker, we compute an attractiveness score using data and insights provided from GS and stock metric information from IEXcloud Api. Given a list of stocks, we can give a recommendation of which stock or stocks to invest in.

How we built it

We constructed a rich dataset by joining GS Marquee data with data pulled from the IEXcloud Api. We determined relevant features to train a prediction model with. For the ML model, we iterated over a few different regression models and settled on a multilayer perceptron regressor. Given the limited number of companies in our dataset (94 companies), we trained with 4-fold cross validation to derive the most learning and avoid overfitting.

Challenges we ran into

We are currently attempting to migrate this into a web app, but we are having difficulty migrating the model into the Django framework

Accomplishments that we're proud of

We were able to train a model with our limited dataset and make informed attractiveness predictions based on stocks that our model had not seen before.

What's next for goldmanHachs

We want to develop this concept into a full-fledged webapp that would provide stock recommendations to the average investor. The attractiveness score metric is a very simple yet informed way to present ranked recommendations to the user. Should the user want to dive deeper into the stock financials, they will also be presented with these numbers as well as resources for future research.

Built With

• django
• google-colab
• iex-cloud-api
• jupyter
• marquee-api
• python
• scikit-learn