This dataset measures how often 216 U.S. public companies file with the SEC,
broken down by form type. It measures filing activity — not company size, revenue, or health.
Data source
Every figure comes from the SEC's own submissions API at data.sec.gov/submissions/CIK##########.json.
EDGAR data is a work of the U.S. federal government and carries no copyright restrictions. The derived tables and
text on this site may be reused freely with attribution.
What exactly is counted
Each company's filings as returned in the filings.recent array of the submissions endpoint.
"Insider forms" means Form 3, Form 4 and Form 5 — statements of changes in beneficial ownership.
"Filings per year" = filings in window ÷ length of the window in years, computed from the earliest and
latest filing dates actually present in that window.
The single most important limitation: windows differ in length
The submissions endpoint does not return a fixed number of filings. In this snapshot the window ranges
from under one year to more than eight years depending on the filer: JPMorgan Chase returns 26,293 filings
covering about one year, while some companies return ~1,000 filings covering eight years.
Consequence: raw filing totals are not comparable between companies. That is why the ranking
uses filings per year, and why every page publishes its own window dates. Always read a count together with
its window.
Which entities are included
Only CIKs that have filed at least one periodic report (10-K, 10-Q, 8-K, 20-F, 40-F or 6-K) are treated as
operating companies. This deliberately excludes natural persons who appear as reporting owners — individual
insiders file Forms 3/4/5 but never 10-Ks — as well as certain funds and trusts. Entities failing this test are
dropped from the dataset entirely rather than shown with misleading numbers.
Known limitations
Sampling bias. The company list is a curated set of large, well-known filers, not a random sample.
Aggregate percentages describe this sample only and should not be read as market-wide figures.
Recent-window bias. Filings older than the recent window are not counted, so companies with a long
history and low recent activity will look smaller than they are.
No corporate-event adjustment. Mergers, ticker changes, deregistrations and name changes are not
normalised. Former names are reported where EDGAR provides them.
Corporate grouping is not deduplicated. A parent and its subsidiaries appear as separate CIKs.
Not investment advice. Nothing here is a recommendation, and no figure is audited.
Reproduce any number on this site
Fetch one file, e.g. https://data.sec.gov/submissions/CIK0000320193.json, then count the
form array inside filings.recent and read the first and last entries of
filingDate for the window. Those two steps reproduce every figure published here.
Corrections
If a figure here disagrees with EDGAR, EDGAR is correct. Each page links to the exact source file used.