Inspiration

We believe that everyone deserves the opportunity to build wealth and secure their financial future, regardless of their background. Through a cutting-edge AI-driven stock prediction model and easy-to-understand resources, we aim to simplify investing and make financial decision-making more approachable.

What it does

The central feature of our platform is our stock price simulator, which attempts to forecast the prices of a given stock over the next seven days. We also have sections of our website dedicated to educating people on stock market basics and resources for financial literacy.

How we built it

The stock simulation begins by retrieving publicly-accessible financial data from Yahoo! Finance. Then, 60-120 days of historical stock prices are stored in a Cloud-based database. We take that historical data and use it to train a neural network model. After training, our app allows you to simulate the stock price for any company listed on the S&P 500. Our tech stack includes: -Bootstrap and JavaScript for the frontend -Flask, tensorflow, and keras for the backend and ML model -MongoDB Atlas for the database

Challenges we ran into

HTML/CSS formatting, lack of sleep, and initial issues setting up our MongoDB Atlas database and connecting to and working with it in Python.

Accomplishments that we're proud of

Producing a working AI model and functional web app in a 24 hour window, working effectively as a team on a tight schedule with different skill sets and technical knowledge.

What we learned

Several of us worked in areas we had little experience in, learning front-end development (HTML,CSS, JavaScript), back-end development (Python and MongoDB Atlas), in addition to refining our data processing and Machine Learning skills.

What's next for Diversify Stock Prediction

To make this project more industry standard, we have to examine other causes for stock prices beyond just historical prices. As stocks fluctuate due to a mix of government action, consumer trust, global events, and other unseen factors, we must train additional models that handle qualitative data. As this process would likely take months, if not years, this remains an idealized goal. In the interim, we can expand the financial education resources hosted on the site.

Built With

Share this project:

Updates