A Usability Study on Data Provenance Visualization Approaches

Abstract

Data visualization becomes an arising need to the end-users to assess, analyze, and reveal the data's insights, especially in the Big Data age. The amount of data generated by platforms and systems over the world increases rapidly. As an illustration, in Social Media and e-Science, domain data increases at a tremendous amount. Data provenance is a sort of metadata on the origin and life cycle of data is the issue of both domains for data analysis. Therefore, this requires further researches to ensure better understanding and efficient data interpretation. Our former comprehensive study has addressed several visualization approaches to meet these needs, such as data summarization and graph comparison in the e-Science domain. Proposed approaches implemented into a visualization tool to illustrate their applicability for the end-users. One of the objections was to involve end-users in the visualization process as much as possible. In this study, we present a detailed user study to assess our implemented visualization environment concerning Human Machine Interface (HMI) from the end-user perspective. Implemented visualization approaches are evaluated using task-level satisfaction and general-level transaction user surveys with well-defined study cases. In addition, System Usability Scale metrics are applied to measure the SUS score of the overall visualization system. The user study experiments proof that the proposed visualization approaches are applicable, and processing overhead is significantly low by facilitating the visualization system.