Description
Some observations were made on the quality of x_analytics_map[wprov]
Data shows many malformed/extended variants of wprov values such as:
yicw1%D8...
rarw1/static/...
sfla1https://...
sfti1%23…
This indicates wprov is often contaminated by appended text/URL fragments, which likely causes undercounting when matching exact values only (especially for external codes).
In this ticket, we need to :
- investigate how much contaminated wprov data we have.
- perform sanitization on the data.