The stock market is basically on sale, and we haven’t seen an opportunity like this since 2008 when the last major recession happened. The coronavirus pandemic is far from over, soif you have a long-term time horizon and you are able to sustain yourself and your family during this period, Invest now while you have decades to recover, financially speaking
But how can we leverage data to help us decide what stocks to buy?
That’s where we come in.
What it does
Stock Patterns 'R' Us is a web platform that dynamically:
- Scans for whether stock movement matches candlestick pattern of your choosing
- Indicates whether the price is likely to rise or fall
- Sifts through available news headlines and analyzes latest sentiment
You might ask why I chose to use candlestick patterns.
First, Candlestick patterns are technical trading tools that investors have been using for centuries to predict price direction. While some patterns can predict more accurately than others and everything is context-based, there’s a reason why the practice exists today.
Second, while candlestick charts basicaly show the same information as bar charts, they do so in in a different way. Candlestick charts are more visual, due to the color coding of the price bars and thicker real bodies, which are better at highlighting the difference between the open and the close.
How I built it
I used Python Flask, Jquery, and html for the front and back end, as well as CSS and Bootstrap for styling. Beautiful Soup was used to scrape news headlines and Natural Language Toolkit was used to analyze sentiments. I also leveraged Alpha Vantage API to obtain up to date stock data, charts, and news articles from stock screener site finviz.
Challenges I ran into
Yahoo Finance API initially did not function for me due to being deprecated or restructured, so I turned to Alpha Vantage API, which gives me the data I want but also has a standard API call frequency of 5 calls per minute and 500 calls per day limit for free users.
Recording was also difficult, as my audio didn't work with the standard windows recorder. And the free video recording and editing apps I used gave me a lot of trouble this time around — I spent 3 hours on this, and this was the best result.
Teammates dropping is a growing issue with online hackathons, too. While this might make this entry ineligible, I don't know what else I could've done.
Accomplishments that I'm proud of
I worked on the back end!!!!! A major step for this front end
What I learned
I learned a bit of looping, a bit of webscraping, a bit of sentiment analysis. This has been a really satisfying project with many uses.
What's next for Stock Patterns 'R' Us
Add more analyses - output a way to know what patterns are available so that I don't always get no returns.