The price of a stock relies on a vast amount of data. During boilermake, a hackathon at purdue, we developed an application that web scrapes hundreds of news articles relevant to a company and projects a sentiment analysis tool onto the data to engineer a new dimension to stock analysis. Very few tools out there do what this application does, and the ones that do exist lack foundamental features that an investment banking analyst would be interested in. Because of the lack of tools in this market, we decided to build an application that would allow a user to investigate this potential correlation even further. More specifically, we decided to focus on public sentiment analysis the company and the CEO vs stock volume.
We plan on our application working with people who are financial analysts, as well, people who are interested in finance. As a hackathon team, we focused on ensuring our features were well established, instead of throwing in empty features after empty features. Our application web scrapes, in live time, 100 news articles relevant to the company inputted by the user. These news articles are then analyzed via a sentiment analysis tool and given a standardized score. Furthermore, we focused on developing a clean and easy to use interface. The interface is very clear and provides the top 5 most negative and positive news articles, as well as recent articles about the website and a chart that makes it easy to compare trends between public sentiment and volume. In the future, we hope to implement concurrency into our application by using a npm library known as cluster to help improve the scalability of analyzing large data sets of news articles.