Three-Component Visual Summary: a Design to Support Casual Experts in Making Data-Driven Decisions

Calvin Yau, Purdue University

Abstract

Recent advancements in data-collecting technologies have posed new opportunities and challenges to making data-driven decisions. While visual analytics can be a powerful tool for exploring large datasets and extracting relevant insights to support data-driven decisions, many decision-makers lack the time or the technical expertise to utilize visual analytics effectively. It is more common for data analysts to explore data through visual analytics and report their findings to the decision-makers. However, the communication gap between data analysts and decision-makers limits the decision-maker’s ability to make optimal data-driven decisions. I present a Three-Component Visual Summary to allow accurate and efficient extraction of insights relevant to the decisions and provide context to validate the insights retrieved. The Three-Component Visual Summary design creates visual summaries by combining visual representations of representative data, analytical highlights, and the data envelope. This design incorporates a high-level summary, the relevant analytical insights, and detailed explorations into one coherent visual representation which addresses the potential training gaps and limited available time for visual analytics. I demonstrate how the design can be applied to four major data types commonly used in commercial visual analytics tools. The evaluations prove the design allows more accurate and efficient knowledge retrieval and a more comprehensive understanding of the data and of the insights generated, making it more accessible to decision-makers that are casual experts. Finally, I summarize the insights gained from the design process and the feedback received, and provide a list of recommendations for designing a Three-Component Visual Summary.

Degree

Ph.D.

Advisors

Ebert, Purdue University.

Subject Area

Design|Web Studies

COinS