Inspiration
The stock market is a bunch of mumbo jumbo money stuff. As a investor who lost $10 with GameStop stock, I can fully proclaim that I don't know what I'm doing. Why does the line go up? Why does it do down? I don't know, but I want to know why. By web scrapping articles from the internet and using sentiment analysis, we can see if articles on the internet have a correlation for stock prices.
What it does
This is a website that takes 3 inputs, a stock, a start date, and an end date. Then it analyses the percent change between those 2 dates and gives a sentiment score between -1 and 1. -1 is negative, 1 is positive, and 0 is neutral.
How we built it
We used python and Flask as a base for the application. We used Yahoo Finance API's to gather information about a stock, and Aylien newspaper API to analyzes articles about the company. For sentiment analysis, we used information from the newspaper API.
Challenges we ran into
Finding a way to find articles to scrape. It's very hard to set up search API's and get an API key without inputting a credit card. Additionally, for sentiment analysis, we wanted to average out all pre-made sentiment analysis of the articles from the newspaper API, and it was hard to figure out an algorithm to do this. We settled on take the total positive sentiment minus the negative, then divide it all by the total scores of the negative, positive, and neutral scores. (pos - neg) / (pos + neg + neu)
Accomplishments that we're proud of
This application works, which in itself is awesome.
I managed to find an API that scrapes articles, and I actually read the documentation to use it.
Figuring out a good way to average the sentiment analysis scores from the articles
What we learned
How to be flexible when trying to find an API and how to change the project structure to suit the technologies that are available.
The usefulness of reading documentation.
What's next
Analyze more articles, display more stock data, show data from articles, basic optimizations to improve performance.
Log in or sign up for Devpost to join the conversation.