Data Quality

A lot of data quality issues don't tend to manifest themselves immediately. So how do you go back and address what those root causes might have been or is that analysis even worth it? It's really important to treat data quality as a real time issue because realistically, it just takes so much energy to try to fix stuff in the past basically. And then you can kind of get ahead by having humans in the loop to fix issues in the past. This is a very good point too because often, once the data is broken you actually can't fixHistoric missing data that someone got mangled and it's only if you're monitoring real time and trying to fix it tries your machine learning model