Projects

These are some of the projects I've worked on, with many more unfinished ones in the works:

As with maybe many students studying computer science, the idea of algorithmic trading crossed my mind as soon as I lost some money investing normally. I decided at the time that the issue had been a lack of rigorous and automated execution and so began my attempts at trading with an algorithm. My first system was quite simple:

• Grab market data every day from Polygon.io for current S&P 500 stocks.
• Search for stocks whose 50-day SMA have just crossed their 200-day SMA.
• Immediately execute market buy orders via Alpaca API - paper trading for now.

While I definitely didn't know what I was doing initially, it opened my eyes to how complex algorithmic trading could be, and that complexity was what captivated me. As a result, one of my good friends Jaiman Parekh, and I have continued on attempting to build a system. While I initially focused on an execution platform, he worked on a backtesting suite that can be found here. At the moment the rest of the repository is still private but I am hoping to make what we have so far public soon.

I've long been very concerned with making the most of my time. I once saw a chart that displayed how a person's perception of time changes as they age:

and this has long stuck with me. I figured that if my perception of time passing might change, it might be important for me to see where it was going. I ended up using Toggl, an app originally designed to track hours for employees that I repurposed to track my time. Then I wrote a small program that sends me a weekly summary of my time distribution across different categories in the form of an email. Here's a bit more about how it works:

• Track all my time with the toggl app. Every day.
• Use the Toggl API to load the data into a Pandas dataframe.
• Create graphics on the time distributions
• Send an email with the graphics via Mailgun API

I ended up hosting the whole project on a DigitalOcean droplet. I set up a weekly cronjob to send the emails every Sunday at midnight, and I've been getting my weekly time analytics ever since.