Meme Stock Watchlist

Notification email
Dynamic site (live)
Site demo
Gemini analysis demo

Inspiration

We've noticed over the last few years that there's been a major shift in financial markets, where retail traders have begun leading the charge on momentum for certain "meme" stocks. The boom and bust cycle is typically very short and some friends/acquaintences of ours have tried, and sometimes unfortunately failed, at getting into the cycle in time, leading to missed profit and sometimes big losses.

What it does

Our project is composed of a two-filter solution: it first searches popular internet forums for activity on public equity tickers and creates a composite score which must pass a threshold for the ticker to move to the next step, where the ticker is matched with elements of its financial info such as volume to determine if the online posting activity has actually translated to early signs of market buy-in. From there, successful tickers that have passed through both filters will be sent as a notification to subscribed accounts, and the ticker with relevant financial info and general context (i.e. what the company does) is posted onto a public dashboard.

How we built it

We wrote the backend in Python using the packages Pandas, integrated Gemini API to provide stock context, and our frontend used React. We also used Reddit API to track internet activity for the first filter, and used various statistical measures such as Z-score on the volume of the stock (most recent tick compared to last 50 days of volume to create a baseline) to set up the second filter.

Challenges we ran into

-Integrating front end with back end (we remained with a static front-end until close to the deadline and had to quickly route our statistical variables and tickers correctly) -Marge conflicts and Git control (we were all pretty new to collaboration projects on Git) -Early project management/flow of tasks -Reddit API limits forced us to only use the most recent 100 posts for our time window -X API free tier is very limited, and paid tiers are out of our reach, so one big resource for online stockwatching communities was inaccessible

Accomplishments that we're proud of

-Our filtration system for the stock tickers is something we're proud of, and we see a genuine use case for this for people who aren't persistently looped into memestock communities but still want to maintain a live memestock watchlist for more trading ideas.

What we learned

-Early project management and flow of backend to frontend as well as order of tasks is critical! -Using Git collaboratively and working as a team to code together one cohesive product -API calls, limits, and use cases

What's next for Meme Stock Watchlist

-Continuation of adjustment of the two-tiered filter system (including more backtesting on various stocks to see how well it avoids false positives/negatives) -Using a wider variety of online forums and communities to gauge how the internet feels about certain stocks in the present moment -Usage of more technical indicators beyond volume and close price to finetune which stocks are primed to see large bullish activity within the next 24h time window