Effectiveness of ChatGPT as a team member in data science project

Publication Type

Conference Proceeding Article

Publication Date

5-2026

Abstract

We studied the effectiveness of ChatGPT as a team member in student groups to complete a data science project from idea generation to deployment, following a modified 7-stage CRISP-DM process. In general, the student groups rated ChatGPT to be effective in terms of the quality, accuracy and relevance of the responses generated. ChatGPT was most effective in Final Report and Data Description stages, and less effective in Data Analysis and Evaluation stages. The student groups were most willing to use ChatGPT in future for Idea Generation and Problem Understanding stages, even when these two stages did not have the highest ratings. Our qualitative analysis for these two stages revealed some interesting insights.

Keywords

ChatGPT, Team Member, Data Science Project, Modified CRISP-DM

Discipline

Artificial Intelligence and Robotics | Databases and Information Systems

Research Areas

Data Science and Engineering

Publication

Proceedings of the 18th International Conference on Computer Supported Education, Benidorm, Spain, 2026 May 18-20

Volume

1

First Page

75

Last Page

83

ISBN

9789897588334

Publisher

SCITEPRESS

City or Country

Portugal

This document is currently not available here.

Share

COinS