Paper Number

ECIS2026-2285

Paper Type

CRP

Abstract

With the prospect of improving organizational performance through advanced analytical methods, Data Science (DS) has become a key driver of many modern businesses. While DS process models (DSPMs) support DS initiatives, their practical adoption remains limited, and project failure rates are high. Furthermore, contemporary DSPMs rarely address the rapid advancements in Artificial Intelligence and Machine Learning (ML). ML Operations (MLOps) has emerged to guide the development and operationalization of ML systems. This paper introduces an MLOps-based DSPM (MLOps-DSPM), constructed from a systematic synthesis of 34 design requirements and formalized using established modeling notations. The design is evaluated through interviews with twelve experts. Future research will apply the model in practice to further validate its effectiveness.

Share

COinS
 
Jun 14th, 12:00 AM

Integrating Engineering Practices Into Data Science: An Mlops-Based Process Model

With the prospect of improving organizational performance through advanced analytical methods, Data Science (DS) has become a key driver of many modern businesses. While DS process models (DSPMs) support DS initiatives, their practical adoption remains limited, and project failure rates are high. Furthermore, contemporary DSPMs rarely address the rapid advancements in Artificial Intelligence and Machine Learning (ML). ML Operations (MLOps) has emerged to guide the development and operationalization of ML systems. This paper introduces an MLOps-based DSPM (MLOps-DSPM), constructed from a systematic synthesis of 34 design requirements and formalized using established modeling notations. The design is evaluated through interviews with twelve experts. Future research will apply the model in practice to further validate its effectiveness.