Discrepancy in matching records to GPS vs number of records
nickrsan opened this issue · 1 comments
nickrsan commented
~430ish records match with GPS vs 477 actual records in the test set. How is this handled in the existing protocol? I mentioned this in the questions for John, but also trying to flag it here - it's a failing test, but I think it's failing for a reason that may need evaluation. Is it OK to discard the 5% of data, or do we need to do a closest match on the GPS points for the ones that didn't have a match?
nickrsan commented
Addressed in another issue