Stocks and Strings

• 

  Dictionary Processing in Python

Inspiration: Providing stock analysis in an era of information saturation (where it's almost impossible/has an extremely high opportunity cost to make people manually read articles and track down words about a company and their stock correlations as an example), and finding cool correlations with words and their effect or correlation on the market.

What it Does: Specifically, analyze correlations between word-usage on news articles reporting on Facebook to Facebook Stock. Broadly, analyze correlations between word-usage on news articles reporting on X company to X stock. ("X" holds multiple possible combinations, from broad areas - i.e. technology, to specific companies/technologies - i.e. Facebook or Litecoin)

Built: 1.) Getting 'word' data from news websites (News API)

2.) Parsing Words & Numbering their frequencies (With more time, phrases, combinatorics of words in their phrases could be added)

3.) IEXCloud; ALPHAvantage APIs to evaluate the U.S. Stock Exchange, specifically for Facebook

4.) Using #2 and #3, creating mathematical models to map keywords to their specific correlations in stock prices

Challenges: 1.) Aesthetics 2.) Time Crunch + complexity of the program - universally functioning aesthetic website (using JavaScript and HTML for the front-end, although the back-end works perfectly)

Links: 1.) Program -- Github -- Files Are Here 2.) Presentation -- Google Slides

Learned: Advanced coding in Python, importing APIs, JavaScript, Stocks/Word correlations

What's next: Increasing API data and making the models fit better! Perhaps modeling a predictive neural network to take the words from the model and turning it into a stock-trading software if the correlation is high enough.

Built With

• datamining
• javascript
• natural-language-processing
• python