TeX 53.6%
Python 46.2%
1# DATA_DICTIONARY.md23Every constructed variable: name, formula, timing, units, availability.4Raw fields first; derived features appended as they are implemented (Stage D+).56## Raw fields (per bar, source: mito0o852/OHLCV-1m @ 7763284)78| Field | Type | Units | Timing | Notes |9|---|---|---|---|---|10| `timestamp` | Datetime(ns, UTC) | — | **start** of the bar minute | verified UTC; convert to America/New_York for session logic |11| `open` | f64 | USD | first trade in minute | raw, NOT split-adjusted |12| `high` | f64 | USD | max trade price in minute | raw |13| `low` | f64 | f64 | min trade price in minute | raw |14| `close` | f64 | USD | last trade in minute | raw |15| `volume` | f64 | shares | sum within minute | integer-valued; venue scope unknown |16| `ticker` | str | — | contemporaneous symbol | NOT a permanent identifier; reuse possible |1718**Key property:** a missing (ticker, minute) row means *no reported trade*, not zero return.19There are no zero-volume rows in the source.2021## Derived variables2223*(to be added with formulas, lag discipline, and config references as Stage D/E code lands;24see `config/shock_definitions.yaml` for planned shock and horizon definitions)*2526## Clean minute panel (`data/processed/minute/year=YYYY/month=MM/data.parquet`)2728| Field | Type | Definition |29|---|---|---|30| ticker | str | contemporaneous symbol (identity unit = spell, see stage C) |31| timestamp | Datetime(ns, UTC) | bar-start minute, deduplicated key with ticker |32| open/high/low/close | f64 | USD, unadjusted; duplicates merged first/max/min/last |33| volume | f64 | shares, duplicates summed |34| date | Date | ET trading date |35| mod | i32 | minute of day ET (570 = 09:30) |36| session | str | pre / rth / post; rth truncated at true XNYS close (early closes) |37| flag_ohlc | bool | Low>min(O,C) or High<max(O,C) or Low>High — flagged, kept |38| flag_split_window | bool | within ±1 trading day of a high-confidence split event |3940## Universe table (`data/metadata/universe_monthly.parquet`)4142Membership for month m computed from month m−1 only (no look-ahead):43days, med_close, med_bars, dollar_vol (m−1 aggregates); eligible (screens);44top500/top1000/top3000 (lagged dollar-volume ranks); flag_noncommon (ticker heuristic).4546## Split events (`outputs/diagnostics/stage_d/split_events.parquet`)4748ticker, date, overnight_ratio, event ("split:R"/"reverse_split:R"), prev_close,49rth_first_open. Windows in split_windows.parquet (ticker, date).50