Applying MLOps to a Data Visualization System: A Case Study

Abstract

This thesis investigates the integration of Machine Learning Operations (MLOps) in the development of a data visualization system, aiming to assess its impact on data quality, code quality, and model quality. The study addressed two primary research questions: (1) To what extent can MLOps be helpful in a data visualization system? and (2) Which best practices can be derived from applying MLOps to such a visualization system? To address these questions, a case study was conducted to investigate if there was an improvement when introducing MLOps to the system. Through a combination of experimental and observational work, the research analysed MLOps improvement from both quantitative and qualitative aspects, and identified best practices from the development of the case study. These findings contributed to bridging the gap between different MLOps applications in data visualization field, and served as a successful example for practitioners, developers, and researchers in the intersection of MLOps and data visualization.