The statistics can give great insights into unusual, suspicious and "unorganic" or unnatural looking activity that can't be as easily noticed when reviewing individual or (small) groups of edits.

There have been and still are many cases of incorrect edits, sometimes so massive that they need to be noticed as soon as possible to keep potential cleanup work as low as possible.

See for example misuse of Soundtrack work type:
https://community.metabrainz.org/t/what-is-a-soundtrack/167398/46

Or the the release country issue with script assisted digital releases

Other examples:

• Recording-Release “samples material” relationships
• Event-Release group “launch event” relationships
• Artist-Release “mix-DJ” relationships
• Videos (April 2021 was me, see also IRC logs)
• Works with attribute “GEMA ID” (These and other work ID are script assisted)
• Abstentions
  • Approvals (script assisted)
• Area-Recording “arranged in” relationships (large spike in December 2023 done by a single editor)
• Artist collections (Big spike near end of year 2023)