Match Score Over Time

Hi ​@rfeith, Excuse me for the lengthy wait. I hope the below notes below are helpful. 

What you can and can’t retrieve from Data Hub for match score “history”

Historical match score changes (a time series)

I don’t have any documentation in the provided sources that Data Hub stores an event-by-event history of match score value changes (for a given candidate on a given requisition) that you can query back as a timeline.

A “last updated” timestamp for the candidate–requisition match score

In the Data Hub data dictionary, the TA hiring outcomes / application-sourcing style datasets include both a match score field and an “application last stage timestamp” field:

• match_score (or application_match_score) is the match score on the application. 
• application_last_stage_ts is the timestamp of the last update in the application process (i.e., last stage update).

However, none of the provided docs explicitly state that application_last_stage_ts equals “the last time the match score was recomputed/updated.” It’s documented as stage/update timing, not match-score recompute timing.

QuickSight datasets and Redshift tables data dictionary

What you can do in practice 

1) If you just need a timestamp that likely corresponds to when the exported snapshot changed

Use application_last_stage_ts as the best available “last updated” proxy in the extracted data, because it is the explicit timestamp you have tied to the application record’s latest update.

2) If you’re observing scores change over time for prior-year reqs

Be aware that exported/warehouse-style data can represent a snapshot at a given generation time, and may not match what the UI shows “right now.” In at least one documented investigation, Eightfold indicated warehouse/DWH data is a snapshot as of when the warehouse data was generated, while the product UI can show a more real-time value—so some inconsistency is expected.