Unique records in the raw_data staging tables, before dbt. Each entity type uses a different deduplication key.
Row counts from the dbt-managed final tables in social_data_alfa. These are the records that passed all dbt transformations.
For each entity: raw LEFT JOIN final on business key, counting rows where the final side is NULL. Results are capped at 500 per entity — the true gap count may be higher.
Each entity: loaded ÷ raw × 100. Then the mean of all 0 entity rates. Note: every entity is weighted equally, so Keywords at 0.5% pulls the average down as much as Posts at 100%.
No gaps detected.
No gaps detected.
No gaps detected.
No gaps detected.
No gaps detected.