Datasets:

Arimancy
/

cftc-cot-weekly

license: cc-by-4.0
pretty_name: CFTC Commitments of Traders Weekly Panel
tags:
  - finance
  - economics
  - commitments-of-traders
  - cftc
  - futures
  - positioning
  - time-series
size_categories:
  - 1M<n<10M
source_datasets:
  - original
configs:
  - config_name: cot_panel_long
    data_files:
      - split: train
        path: cot_panel_long.parquet
  - config_name: cot_features_weekly
    data_files:
      - split: train
        path: cot_features_weekly.parquet
  - config_name: markets_reference
    data_files:
      - split: train
        path: markets_reference.parquet
  - config_name: release_anomaly_calendar
    data_files:
      - split: train
        path: release_anomaly_calendar.parquet

CFTC Commitments of Traders weekly panel

Every CFTC Commitments of Traders report family in one tidy, model ready weekly panel: harmonized positions, net positioning and COT index features, a market reference map, and documented release provenance, from 1986 to last Friday, in Parquet and CSV.

Dataset structure

Four tables, each its own named config (different schemas, never concatenated):

cot_panel_long: the tidy long panel, one row per (report_date, contract code, report family, basis, trader category).
cot_features_weekly: the model ready wide table, one row per report, with open interest and its change, concentration ratios (plus a source anomaly flag), and per category net position, net share of open interest, change from the prior report, COT index over 52 and 156 reports, and a 156 report z score. Windows partition on market_key, the reviewed contract coherence surrogate.
markets_reference: one row per contract code and family.
release_anomaly_calendar: documented publication and methodology events.

from datasets import load_dataset

panel = load_dataset("Arimancy/cftc-cot-weekly", "cot_panel_long")
feats = load_dataset("Arimancy/cftc-cot-weekly", "cot_features_weekly")

Load a table by config name. Full column documentation, category definitions, the cadence by year table, the dropped columns appendix, and the leakage note are in the included data_dictionary.md.

Why this exists

The CFTC publishes COT across several report families with different start dates, schemas, and filename conventions, and its own FAQ notes the data has no primary key. This dataset reconciles four decades of schema drift into one schema and makes the release history explicit. The October to December 2025 appropriations lapse, when publication stopped for six weeks and was backfilled in December, is a reminder that when a number was released is part of the data.

Caveats

Positioning data only. No prices. The market reference map provides join keys; users bring their own price data under their own licenses.
Trader classification is self reported business purpose on CFTC Form 40, subject to CFTC reclassification. Categories are regulatory classifications, not strategy labels.
report_date is the as of date published by the CFTC, usually Tuesday close; 163 dates are not Tuesdays (the early non weekly history plus holiday shifts). The report_weekday and is_standard_tuesday columns make the exceptions filterable. Releases are usually Friday 3:30 pm ET, shift around holidays, and are irregular during documented events. Use the release anomaly calendar; never assume report_date plus three days.
The October to December 2025 window: report_dates are complete after the chronological backfill, but original release timing was interrupted.
Early history cadence is not weekly; see the cadence by year table.
Combined basis converts options to futures equivalents using exchange supplied deltas (CFTC methodology).
Nonreportable is a derived residual (open interest minus reportable positions); trader counts do not exist for it.
Contract consolidation methodology has changed over time per CFTC announcements; the panel reports what the CFTC published for each date.
CFTC reuses contract market codes. Where a code splices two different underlying contracts, the features table separates them with a market_key so window features never blend them.

Attribution and license

Source: CFTC Commitments of Traders reports (US Government work, public domain). This product is not endorsed by or affiliated with the CFTC. Published by Arimancy LLC under CC BY 4.0; the license text and the attribution target ship in the package as LICENSE.txt. Dataset page: arimancy.com/datasets/cftc-commitments-of-traders-weekly.