Verification Checklist -- P2

Storage data

• Total storage rows: 12219
• Elevator A rows: 6947
• Elevator B rows: 5272
• Per-month counts (Elevator A): jan=1209, feb=1073, mar=1185, apr=1165, may=1162, jun=1153
• Per-month counts (Elevator B): jan=902, feb=827, mar=889, apr=865, may=921, jun=868

Weather extraction

• Expected row count: 181 (one row per day, January 1 to June 30, 2025)
• API total: The Open-Meteo API reports the total number of days in the response. The extracted count must match.
• Date range: 2025-01-01 to 2025-06-30

Staging verification

• stg_storage row count must match raw storage count exactly: 12219
• stg_weather row count must match raw weather count exactly: 181
• All bin IDs in stg_storage must follow standardized format (elevator prefix included)
• All weather timestamps in stg_weather must be in UTC+6 (Kazakhstan time)

Mart verification

• Expected mart row count: This depends on the grain decision (one row per bin per day with weather data for the same date). The count should equal the number of storage readings that have a matching weather date.

• Spoilage-weather correlation spot-checks:

  • Date: 2025-01-08, Bin: A-025, Grain: wheat, Moisture: 10.7%

    • Weather on this date (from Open-Meteo for Kostanay): temperature should show below -15C min, humidity above 80%
    • The mart record should show both the storage reading AND the weather conditions joined on date

  • Date: 2025-01-09, Bin: A-025, Grain: wheat, Moisture: 13.1%

    • Weather on this date (from Open-Meteo for Kostanay): temperature should show below -15C min, humidity above 80%
    • The mart record should show both the storage reading AND the weather conditions joined on date

  • Date: 2025-01-15, Bin: A-013, Grain: wheat, Moisture: 12.2%

    • Weather on this date (from Open-Meteo for Kostanay): temperature should show below -15C min, humidity above 80%
    • The mart record should show both the storage reading AND the weather conditions joined on date

Idempotency

• Running the extraction twice produces the same row count: 181
• Running the full pipeline twice produces the same mart row count
• No duplicate records after re-run

Summary counts