Ten simple rules for humane data science

Abstract

The capabilities of data science can help us see hidden patterns, customize services, and advance biomedicine and science. As data science permeates industry and academia, a question that often arises is: How can we use these capabilities to genuinely help the world? Considering this question brings to mind terms such as “responsible data science” and “data for good.” Inherent in these terms is the desire that data science improve the human condition-in other words, the desire to undertake humane data science. How can we do this? What follows are 10 simple rules to contemplate. (Note that we are referring to data science in a holistic way. Along with machine learning and AI methods, we include all aspects of the data acquisition, analysis, dissemination, and application pipeline, and their accompanying human and socioeconomic factors).