This repository contains code to conduct quantitative analysis on constituents of the S&P 500. The focus is on conducting automated fundamental analysis of stocks in a manner that is free of survivorship bias.
The ultimate objective is to create investment portfolios that outperform the S&P 500 over the long run - that is, a horizon greater than 5 years.
- A computer able to run docker images. You can, as I do, just use github actions to execute your code.
- An alphavantage API key - you can get a free one, but most of this code only really works well with a 75 calls/minute premium API key.
- An AWS S3 bucket with the relevant permissions to push and pull data. You could use any S3 compatible object store, such as minio.
The code in this repository runs on the docker image generated from the Dockerfile, which is built on top of riazarbi/datasci-gui-minimal. It's pretty simple, and merely adds R and python libs for quantitative financial analysis.
The various scripts in the root of this repository scrape data sources and store them in a structured manner in AWS S3.
The functions in the R subdirectory are helper functions to enable the versioned storage of parquet formatted datasets in AWS S3.
Pretty much all the code in this repo is executed on a periodic basis via github actions. You can see what actions are run in the .github/workflows subdirectory.
At present, this just contains an up to date copy of the wikipedia S&P 500 constituent html page. A github action snaphots the wikipedia page and, if it has changed, commits it to the repo. The reason I do this is so that if, for whatever reason, tidyquant's S&P 500 constituent list stops working I can use this as a fallback after doing a bit of html wrangling. I don't snapshot the tidyquant version because I'm unsure of the licensing implications - but I do version it in S3.