Skip to content

Latest commit

 

History

History
32 lines (21 loc) · 1.77 KB

README.md

File metadata and controls

32 lines (21 loc) · 1.77 KB

Google Trends Bond Futures Modeling

Introduction - 4/3/23

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

Alternative Data- 4/3/23

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.

Rudimentary BackTesting

Backtesting is done using the Backtesting.py package

Beginning Hypothesis of Results - 4/3/23

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

Resources

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/

Credit For beginning equityindex code - Ming Fong and Alexander Yang

Future Credit to Datadunce if I get around to Barchart 1min implementation