Idea_2_Description.md 1.8KB
In 2013, my freshman year of college, I stumbled upon coinbase when a friend of mine was bragging about the return he made from holding bitcoin. The concept of cryptocurrencies let alone investing was foreign to me at the time, but I foolishly bought in not knowing what I was doing. When I started studying computer science a few years ago, I began trying to understand the underlying blockchain technology and why so many were considering it to be revolutionary. Aside from the theory behind it, I was curious how these cryptos were being traded and how to manage my own portfolio. Knowing the space is still incredibly immature and susceptible to manipulation, I wanted to attempt to draw correlations between unconventional data and price movement. From the traders that I follow on twitter, I gathered that many were using a combination of sentimental analysis along with technical analysis to devise their trading strategies. I was very curious as to what “alternative data” I could use in a trading strategy. For my project, I wanted to backtest a strategy using a combination of the following:
Developer Index
- Analyze GitHub commits, forks, pull requests and contributors since GitHub development precedes major development in underlying tech.
ICO Database
- Determine actionable trends in successful ICO’s.
Sentimental Analysis
- Twitter especially has its own crypto-community. Maybe attempt to use their API to identify popularity of tags or sentiment of popular ”crypto” accounts.
- Query crypto related searches on google (using trends.google.com)
Network hash rates(probably too difficult)
- Extract forward price movements from miner power allocation
Evaluation:
- Choose different durations
- Just test bull market
- Just test bear market
- Just test flat market
(Possibly learn how to use quantopian to backtest)