Project Artifacts for the Data Science Lifecycle: A Comprehensive Overview

Through knowledge extraction from data with various methods, Data Science (DS) allows organizations to achieve improvements in performance. The execution of these projects is mainly supported by DS process models such as CRISP-DM. As a high percentage of DS undertakings are failing, revisions to current DS project management practices become necessary. Amongst others, ensuring traceability, reproducibility, and knowledge retention across the project present important success factors in DS projects. Some of the DS process models feature documentation artifacts for this purpose but not comprehensively for the complete DS lifecycle. Accordingly, in this research, existing documentation deliverables for DS are identified and examined by means of a literature review. Based on the established best practices from the process models, the contents of 18 derived project artifacts for DS documentation are synthesized for the DS lifecycle to improve DS project management.