go-weatherstation/docs/rain_data_issues.md

Rain Model Data Issues and Mitigations

This document captures known data-quality issues observed in the rain-model pipeline and the mitigation rules used in code.

Issue Register

Area	Known issue	Mitigation in code/workflow
WS90 rain counter (rain_mm)	Counter resets can produce negative deltas.	rain_inc_raw = diff(rain_mm) then rain_inc = clip(lower=0); reset events tracked as rain_reset.
WS90 rain spikes	Isolated large 5-minute jumps may be sensor/transmission anomalies.	Spikes flagged as rain_spike_5m when increment >= 5.0mm/5m; counts tracked in audit/training report.
Sensor gaps	Missing 5-minute buckets from WS90/barometer ingestion.	Resample to 5-minute grid; barometer interpolated with short limit (limit=3); gap lengths tracked by audit.
Out-of-order arrivals	Late MQTT events can arrive with older ts.	Audit reports out-of-order count by sorting on received_at and checking ts monotonicity.
Duplicate rows	Replays/reconnects can duplicate sensor rows.	Audit reports duplicate counts by (ts, station_id) for WS90 and (ts, source) for barometer.
Forecast sparsity/jitter	Hourly forecast retrieval cadence does not always align with 5-minute features.	Select latest forecast per ts (DISTINCT ON + retrieved_at DESC), resample to 5 minutes, short forward/backfill windows, and clip fc_precip_prob to [0,1].
Local vs UTC day boundary	Daily rainfall resets can look wrong when local timezone is not respected.	Station timezone is configured via site.timezone and used by Wunderground uploader; model training/inference stays UTC-based for split consistency.

Audit Command

Run this regularly and retain JSON reports for comparison:

python scripts/audit_rain_data.py \
  --site home \
  --start "2026-02-01T00:00:00Z" \
  --end "2026-03-03T23:55:00Z" \
  --feature-set "extended" \
  --forecast-model "ecmwf" \
  --out "models/rain_data_audit.json"

Operational Rules

• Treat large jumps in rain_reset_count or rain_spike_5m_count as data-quality incidents.
• If gaps_5m.ws90_max_gap_minutes or gaps_5m.baro_max_gap_minutes exceeds one hour, avoid model refresh until ingestion stabilizes.
• If forecast rows are unexpectedly low for an extended feature run, either fix forecast ingestion first or temporarily fall back to baseline feature set.