We were inspired by the inability to predict stocks when we are looking at their charts.

What it does

StockOverflow predicts the stock price for 15 days in the future, and gives a score for the sentiment of the news articles released about the stock.

How we built it

We built it as a flask web app (python). The sentiment analysis model was a google cloud API, where as the stock price predictor was a model we wrote with tensor flow. We have the site hosted on a google cloud app engine. One can check it out at out link below:

Challenges we ran into

We ran into issues setting up the website and configurations that needed to be made to do so, as we did not have prior experience setting up a site.

Accomplishments that we're proud of

We are proud of the clean look of out site, as well as the integration of the Dino game in our waiting page.

What we learned

We expanded our knowledge within python, and learned flask. We also learned about web scrapping in python.

What's next for StockOverflow?

In the future, StockOverflow could be worked on to improve the accuracy and take into account a larger set of factors for its predictions. The sentiment analysis could be expanded to include comments on various online forms about the companies, rather than just news articles.

+ 4 more
Share this project: