Paper Number
1879
Paper Type
Complete Research Paper
Abstract
Data science (DS) projects have become indispensable across a wide range of industries. Various data science process models (DSPM) have been proposed to provide structure and guidance for the project process. Despite recognizing the efficiency of DSPM, their practical adoption has been limited. However, the increasing frequency of data science projects and the associated repetitive process execution highlight the importance of this gap and justify further investigations. To support the application of DSPM by practitioners and guide further research, a comprehensive overview and systematic categorization of DSPM is necessary. Our research addresses this need by creating a taxonomy of DSPM. We utilized both inductive and deductive research methods to develop a taxonomy that includes three categories, 13 dimensions, and 67 characteristics. The taxonomy is demonstrated using a sample of 35 DSPM from the literature and evaluated through twelve expert interviews with DS experts and researchers.
Recommended Citation
Rösl, Stefan and Schieder, Christian, "A Taxonomy of Data Science Process Models: Insights from Science and Practice" (2024). ECIS 2024 Proceedings. 12.
https://aisel.aisnet.org/ecis2024/track07_busanalytics/track07_busanalytics/12
A Taxonomy of Data Science Process Models: Insights from Science and Practice
Data science (DS) projects have become indispensable across a wide range of industries. Various data science process models (DSPM) have been proposed to provide structure and guidance for the project process. Despite recognizing the efficiency of DSPM, their practical adoption has been limited. However, the increasing frequency of data science projects and the associated repetitive process execution highlight the importance of this gap and justify further investigations. To support the application of DSPM by practitioners and guide further research, a comprehensive overview and systematic categorization of DSPM is necessary. Our research addresses this need by creating a taxonomy of DSPM. We utilized both inductive and deductive research methods to develop a taxonomy that includes three categories, 13 dimensions, and 67 characteristics. The taxonomy is demonstrated using a sample of 35 DSPM from the literature and evaluated through twelve expert interviews with DS experts and researchers.
When commenting on articles, please be friendly, welcoming, respectful and abide by the AIS eLibrary Discussion Thread Code of Conduct posted here.