Systematically analyze SEC filings to feed insights into portfolio and risk management models using Natural language processing (NLP) for a UK based Buy-side Firm
Objective
- Client, a UK Based buy-side firm, was starting to incorporate quantitative overlay to the investment process
- They wanted to implement quantitative process based on a paper called ‘Lazy Prices’ published by Harvard
- Systematically analyze SEC filings of Russell 3,000 firms to feed insights into portfolio and risk management models
Platforms used
- Python, Apache Spark, NLP Algorithms
Client impact
- Strengthened portfolio and risk management models by generating complementary and uncorrelated signals
- Scored firms based on changes in text structure and content, sentiment analysis. Z-scores helped comparison within sectors and broader market
- Used advanced NLP algorithms to undertake period on period analysis and quantify change in structure and content of the text
Questions