Research Background:

With the development of new technologies such as social media and sensing, more data-based user insight methods are emerging. The emergence of Large Language Models (LLMs) and automated insight methods for understanding multidimensional data provides an opportunity to address this problem.

Research Question:

In this research, we mainly focused on those three questions:

  • RQ1: What are the opportunities and challenges that user researchers face when conducting user data insights?
  • RQ2: How can large language models be used to help user researchers gain insight from user data?
  • RQ3: What benefits and challenges can large language models bring to researchers?

Overall Experiment Architecture Diagram:

We first created a five-dimensional (5D) dataset toolkit for scenarios where older adults drink water and take medication, and then recruited 20 researchers experienced in user research to participate in the experiment. We found that utilizing the Data Insight Iteration(New method ) method can boost efficiency, providing tailored, credible, and credible analytical conclusions that align with the researcher’s background. However, it has current technical limitations, handling only a singular data format.

Overall Experiment Architecture Diagram

The Paper's Overall Contributions:

  • Analysis of Pain Points and Designer Needs in Traditional User Research (Overall Pain Points)
  • Introducing a Collaborative Approach to User Data Insights (Data Insight Iteration Process)
  • Empirical Evaluation of the Proposed Processes
  • Development of a Five-Dimensional Data Toolkit

Overall Pain Points

Pain points

Data Insight Iteration Process

Data Insight Iteration Process
Made in Webflow