[danielsf]

Author of the Medium post here...

This point comes up a lot...there's a tension in making the data say something interesting that will get traffic/spread, which might undermine the rigor that goes into real data-analysis/data science/academic work.

IMO, as long as we disclose the source and preface the biases/problems, I'm ok with data that isn't perfect (after all, there's no such thing as a perfect data set).

The lyrical analysis that I did for rappers would never work in academia...the data set wasn't strong enough. But, it was good enough for the Internet as a side-project, and I think that most readers understood the integrity issues with the data (which I also highlighted in the narrative).

But yea, really good points about journalistic standards for coders who write journalism-esque content.