Data audit
| file | SHA-256 | rows (excl. header) | cols | range | missing | dup dates |
|---|---|---|---|---|---|---|
| GDPC1.csv | 1c6c6e9f59e383e20c849dd299175c7a307526c09abdc18f18da9754fc5b3343 (matches README) | 319 | 2 (observation_date, GDPC1) | 1947-01-01 .. 2026-07-01, quarterly, no gaps | 0 | 0 |
| UNRATE.csv | 621ff6fadf50910cb7a21ee20ba610c247cb043b29b4e05c4c1a72fadca2a06c (matches README) | 945 | 2 (observation_date, UNRATE) | 1948-01-01 .. 2026-09-01, monthly, no gaps in dates | 1 (2025-10-01, empty value) | 0 |
Units: GDPC1 billions chained 2017 USD SAAR (only log-differences used, so units/annualisation of the level drop out); UNRATE percent SA. GDPC1 range 2172–24611; UNRATE range 2.5–14.8 — plausible.
Anomalies: UNRATE 2025-10-01 blank (BLS did not publish October 2025 during the federal shutdown). Outside the 1959Q4–2019Q4 window, so irrelevant; handled by dropping out-of-window observations.
Information availability: both files extend beyond 2019Q4 (current vintage, 2026). Post-2019Q4 rows are excluded before any computation. Note the GDPC1 vintage is 2026 (chained 2017 dollars), i.e. revised data, not real-time — acceptable since the task is a historical estimate, but noted as a limitation.