Harvard Data Science Review, 3(3), 2021
Over a half-century ago, Tukey and Wilk (1966) launched a rhetorical broadside against purely formal approaches to data analysis. In the intervening decades, software support for statistical modeling and exploratory visualization has made great strides. Despite these advances - or perhaps because of them - we also find ourselves in the midst of a reckoning on issues of analytic transparency, reliability, and replication. Might the time be right for rapprochement?