Quantopian's community platform is shutting down. Please read this post for more information and download your code.
Back to Community
Online Long/Short Fundamental Equity Strategy

I having been trolling these forums for some time and finally decided to start learning python. I am still unfamiliar with all the nuances and whats capable to do in python, so I was hoping to gather some information and hopefully get pointed in the correct direction.

Experience: Undergrad in Finance, Co-manage a $10m quant HF.

What I want to do is:

At the beginning of every week:
- Find the betas for a bunch of factors using the fundamental properties from market open monday the week before regressed over the returns for that week excluding stocks that had an earnings release for that week.
- Determine the predictability beta of each factor by inserting the recent obtained betas into a exponentially weighted moving average formula - Hope to use kalman filter but thats out of my league
- Run PCA on the top factors that describe some optimal % variability in hopes to minimize the number of factors and correlation between factors
- Long the top percent short the bottom percent
- Rebalance every week

This should be able to be done every 6 months, month, week, day etc. The key is for it to adapt to factor betas as new information is obtained.

It wasnt clear that pipeline was appropriate for this task as pipeline is run every day. I need rankings, but the main glut of it is running the weekly regressions on new information.

I imagine this algo will need a year warmup time so each factor ewma has some data so it isnt to volatile.