Towards Actionable Data Science: Domain Experts as End-Users of Data Science Systems

As a wider range of organizations explore using data science systems, research has given growing attention to the role domain experts. Most this still views systems as centered on development statistical models or algorithms by technical scientists, with experts limited informers. Our paper turns how mediate whether lead action through their situated practices. Drawing ethnographic fieldwork and pilot machine learning project at craft brewery, we identify situations where brewers’ practices led unreliable, incomplete data, unpack such effectiveness activities. Extending in CSCW STS experts’ context, aim inform design that are more actionable for end-users.