For speed, I will scriptkiddye Ming/Fong/Datadunce's existing code to prelim test Momo AND MeanRev trading strats for US Treasury Futures based on Google Trend data for certain selected keywords. After general delta 1 move backtests on daily closes, will implement Relative Value Aspects
Google Trends search volume data = punting US Treasury Futures or index/crypto/equity/future.
Will work on mashing /Datadunce 's barchart code later (https://gist.github.com/cf7ad12df1a3d74fd8fee28e16639419.git).
TrendData is downloaded through the Pytrends module using the Google Trends API. I'll reserve how I got daily granular data for my next employer and their investors/shareholders.
Backtesting is done using the Backtesting.py package
High-Level Personal Bias: Daily close historicals not enough to run real money. Need Extended Hours 1min + Reg Hours 1min + ICE Exchange Interest Rate Swap 1min Data Momo Bias - Will be undwhelming 2015 - 2017 and will work in spurts correlated to recent fed hiking cycle MeanRev Bias - promising enough for me to waste time on this
Papers: https://www.researchgate.net/publication/326503702_Algorithmic_Trading_Systems_Based_on_Google_Trends https://jackdry.com/predicting-realized-volatility-using-google-trends Datadunce Barchart 1min code - (https://gist.github.com/cf7ad12df1a3d74fd8fee28e16639419.git) https://pypi.org/project/yahoofinancials/