13f data parsing

13F Data Parsing

This skill outlines how to process SEC Form 13F data structured as TSV files (COVERPAGE, INFOTABLE, SUMMARYPAGE, etc.).

Requirements

• pandas
• dask (optional for large files)

Common Operations

1. Loading the Data

Use pd.read_csv with sep='\t' to read the files. For large files like INFOTABLE.tsv, you might want to specify dtype or use low_memory=False.

import pandas as pd

coverpage_df = pd.read_csv('COVERPAGE.tsv', sep='\t', low_memory=False)
infotable_df = pd.read_csv('INFOTABLE.tsv', sep='\t', low_memory=False)
summary_df = pd.read_csv('SUMMARYPAGE.tsv', sep='\t', low_memory=False)

2. Finding Fund Accession Number

The COVERPAGE contains the name of the institutional investment manager.

# Assuming 'FILINGMANAGER_NAME' or similar column exists.
# We will use fuzzy matching to find the exact name.

3. Extracting Fund Details (AUM, Number of Holdings)

Once you have the ACCESSION_NUMBER, you can look it up in SUMMARYPAGE or INFOTABLE. AUM can be calculated as the sum of VALUEDOLLAR (or similar column) in INFOTABLE for the specific accession number, or sometimes it's reported in SUMMARYPAGE. Number of holdings can be the count of rows in INFOTABLE for that accession number.

4. Analyzing Changes between Quarters

You can join INFOTABLEs from two quarters on CUSIP for a specific fund to find differences in share count or value.