CRISP-DM Twenty Years Later: From Data Mining Processes to Data Science Trajectories

CRISP-DM(CRoss-Industry Standard Process for Data Mining) has its origins in the second half of nineties and is thus about two decades old. According to many surveys user polls it still de facto standard developing data mining knowledge discovery projects. However, undoubtedly field moved on considerably twenty years, with science now leading term being favoured over mining. In this paper we investigate whether, what contexts, CRISP-DM fit purpose We argue that if project goal-directed process-driven process model view largely holds. On other hand, when projects become more exploratory paths can take varied, a flexible called for. suggest outlines such trajectory-based might look like how be used categorise (goal-directed, or management). examine seven real-life exemplars where activities play an important role compare them against 51 use cases extracted from NIST Big Public Working Group. anticipate categorisation help planning terms time cost characteristics.