Closed pwalsh closed 8 years ago
@rgrp @tryggvib
It would be good to address this properly (spec here, or user story) in this sprint (or towards end, going into next one).
In the results aggregator (https://github.com/okfn/spd-admin/blob/master/spd_admin/tasks.py#L97)
My "algorithm" is simply: start at score of 10, minus 1 per error, down to 0.
Obviously we'll want something a bit smarter: something that perhaps treats structural errors differently to schema errors, or, handles cases for example where one single mistake is repeated over multiple rows.
Briefly addressed in #1, we need a scoring algorithm based on results of data validation.
Can't be done without further discussion.