We were inspired by the recent events in the stock market revolving around the effect of media and social exposure on specific stocks.

What it does

Our app uses Twitter, Reddit, and popular news sources around the globe to determine how much the stocks were talked about in external media.

How we built it

We built this platform through the Shiny framework using the tools in the R language and datasets and APIs scraped from existing algorithms. We utilized data from Gambiste and Yahoo Finances and the Reddit and NewsAnchor APIs.

Challenges we ran into

We had to deal with a bulk of stock data that took an extremely long amount of time to process which resulted in the website slowing down. We also utilized a plethora of APIs and packages that we were initially unfamiliar with.

Accomplishments that we're proud of

We are especially proud of how we solved each of the major challenges by patiently approaching the problem. For instance, we utilized the top 10 stocks according to Gambiste in our analysis to resolve the issue of the slow processing of our website, and we closely looked at the documentations and were able to effectively utilize the foreign APIs and packages to our advantage.

What we learned

Through our experience in implementing unfamiliar APIs and packages, we learned the importance of finding expertise in quickly learning new technical skills as well as the value of being patient with resolving issues within our code.

What's next for Stocklytics

We have already started our next step in pursuing a machine learning algorithm that uses past trends to predict future trends based on the intentionality of mentions in social media platforms. We've already started building training models and implemented machine learning libraries in our code! The next step after that would be to expand the analysis beyond the top 10 stocks according to Gambiste to the majority of existing stocks.

Built With

  • animated-graph
  • css
  • ggplot
  • newsanchor
  • quantmod
  • r
  • reddit
  • rvest
  • shiny
  • tidyverse
Share this project: