Wanted to created a scanner for equity themes that could be moving. A Q3 report by BOA did a great job by coming up with themes which they thought are promising and curated all the counters in it.
https://themealert.nn.r.appspot.com/
This is what I did. PDF articles -> extracted info into CSV -> copy the data into Googlesheet (which updates prices live) -> reads into Streamlit app for visual analysis -> deploy Google App Engine
For the tech people, the only painful part was the manual copying of data into GoogleSheet. 1 friend helped me with it and it still took us ard 3 hrs..
- Streamlit is great for fast prototyping. There are extension for row selection etc but I decided the bare functionalities were good enough for analysis
- Googlesheet lags at the start when the number of counters gets into the thousands. Especially when I am pulling historical prices for it too.
- Deploying to google app engine was very easy. For stateless app like this, google cloud run might be a cheaper option.
Took around 2 afternoons to get the app up & running. Pretty fun given that I have not been doing frontend and app deployment for quite a while