Incomplete companyfacts Data – Proposal to Fill Missing Values from Filings

[Cemlyn]

I've noticed that the companyfacts json files tend to be incomplete for certain key data points, e.g. 'CommonStockSharesOutstanding'. For META this fact is missing, but for NVDA it isn't. This makes comprehensive analysis on all stocks tricky. For example, i'm wanting to calculate valuations for all tickers and compare to their current price. Could there be a use case for functionality where the edgartools first get's data from companyfacts and submissions jsons, and then scans the actual filings for missing data points if they are missing?

I realise this may be outside the scope of this package, but just wondering if anyone else has had this issue?

[Cemlyn]

I'm wondering if I'll get more coverage on missing data points by getting all filings and parsing them into a tabular form myself using the full_text_submission function.

[dgunning]

Both NVDA and META have WeightedAverageNumberOfSharesOutstandingBasic so you can compare on those. You are right, META is missing CommonStockSharesOutstanding which is a point in time figure.

However CommonStockSharesOutstanding is in META's xbrl

c = Company("META")
f = c.latest("10-K")
xb = f.xbrl()
xb.instance.query_facts(concept='us-gaap:CommonStockSharesOutstanding')

[Cemlyn]

Thanks. Looks like extracting from each xbrl filing might be the best way to get full coverage across all tickers. I’ll experiment with that approach and see how consistent it is.

This project is proving to be super useful however - I appreciate it!