EPSViz

If predicted EPS beats wallstreet estimates
If predicted EPS falls within wallstreet estimates
If predicted EPS misses wallstreet estimates
If EPS can't be predicted based on available data
shy chart
frontend not really syncing up... but the magic crystal ball says its going to beat wallstreet estimates this upcoming quarter!

Inspiration

One day, we were watching the stock markets and saw a random company gap up 20% after their earnings. We wondered, could this have been predicted? So we looked at data from Google Trends to see if the company's product's search interest would match well with how their earnings report went. Interestingly, some companies' past earnings report history does correlate with their Google Trends search interest! So we proceeded to build this.

What it does

EPSViz does simple linear regression to visualise and predict how companies could perform for their upcoming earnings.

How we built it

We built the backend using NodeJS, R, Express, Postgres, and Docker. Frontend was built using React.

Challenges we ran into

We spent several hours trying to connect R to NodeJS, but after several hours realised that the library we were using was terrible and switched over to another one which worked.