A Lean Design Thinking Methodology (LDTM) for Machine Learning and Modern Data Projects

Abstract

As data projects become more conventional, increase in the use of information has surpassed the knowledge of how to support individuals/teams that undertake such projects. Leading data mining methodology, CRISP-DM has become limited in managing the requirements of working with recent technologies such as Machine Learning. Resultantly, many have either created their own methods or adopted alternative approaches such as the Design Thinking and Lean Startup innovation strategies. Consequently, this paper proposes a novel software development methodology entitled Lean Design Thinking Methodology (LDTM) to guide the development of modern data projects. LDTM combines the strengths of CRISP-DM with the more innovative Design Thinking and Lean Startup strategies to introduce an approach divided into three stages, comprising of seven steps. This paper concludes on how there is no one correct method, nor is one single approach enough, but together, elements of each approach can unite to help guide data projects forward.