Stochastic behavior of stocks is sometimes heavily based on the sentiments regarding the companies' news. While people can get the general sentiment by browsing multiple news and social media platforms, there is no centralized application for parsing these attitudes.

What it does

Our web application allows users to track changes in sentiments of public companies they are interested in and how that affects risk factors. It analyzes the sentiments of news articles from many reputable news sources and tweets on Twitter about the given company, and returns a positive or negative sentiment index for that company, as well as the risk factor of investing in the company over the past month.

How we built it

Data Collection: Stockastic scrapes information from reputable news sources, such as Associated Press, Bloomberg, and Reuters, using the News API, and scrapes relevant tweets from Twitter using Twitter's Developer API. In addition, this application uses BlackRock's API to analyze the risk of purchasing stocks.

Web Application: For a lightweight web application, we decided to use a webstack solution, specifically the MongoDB-Express-React-Node.js stack. For the front, we also integrated some CSS to format the React components.

The system is based on users, who can then save stocks to their accounts. Going through the saved stocks, the users can scroll through their customized list of stocks to view the sentiment of each corresponding company, as well as the risk factors of buying the stocks.

Challenges we ran into

  • Integrating APIs:We had some issues using the APIs at first, including interpreting the resulting parameters from the response
  • MongoDB issues: had some issues using Schemas in other Schemas
  • Issues with CORS: issues with CORS, ended up bypassing it to focus on other aspects of our project

Accomplishments that we're proud of

  • Scrapping the web with javascript (and not python)
  • Putting together a sentiment analysis algorithm
  • Making a functional MERN Stack application to manage the stocks

What we learned

We learned how sentiment analysis works, including the machine learning algorithms behind it. We also learned about the techniques of developing a web application, specifically the MERN stack (and the associated issues with it).

What's next for Stockastic

  • Portfolio analysis. Knowing the past history of the user's portfolio can help them determine what stocks they would like to buy and sell.
  • Better interface. We were concerned with functionality, so with some more time, we could improve the user interface.
Share this project: