A Data Science Central Community
Ronald Coase died last week. Coase was an economist and a Nobel laureate, not someone you would typically associate with modern data analytics. Still, Coase is noted in the field for coining the phrase, “If you torture data long enough it will confess.”
Coase’s quote, and his career, are reminders that analysis can have repercussions that go beyond the screen and the analysis, and have impacts on the work that we do and the people we serve. It’s not malicious intent that causes people to torture the data until it gives them what they want, but the processes of sampling, querying, and modeling data lead us to conclusions that represent only a selection of the data set. And when we are allowed to choose the path to data discovery, there will inevitably be opportunities for bias, error, and time spent on exploratory dead-ends.