User Experience In Dataset Search

Yihang Zhao,Albert Meroño-Peñuela,Elena Simperl

2024-08-05

Abstract:This research investigates User Experience (UX) issues in dataset search, targeting Google Dataset Search and <a class="link-external link-http" href="http://data.europa.eu" rel="external noopener nofollow">this http URL</a>. It focuses on 6 areas within UX: Initial Interaction, Search Process, Dataset Exploration, Filtering and Sorting, Dataset Actions, and Assistance and Feedback. The evaluation method combines 'The Pandemic Puzzle' user task, think-aloud methods, and demographic and post-task questionnaires. 29 strengths and 63 weaknesses were collected from 19 participants involved in roles within technology firm or academia. While certain insights are specific to particular platforms, most are derived from features commonly observed in dataset search platforms across a variety of fields, implying that our findings are broadly applicable. Observations from commonly found features in dataset search platforms across various fields have led to the development of 10 new design prototypes. Unlike literature retrieval, dataset retrieval involves a significant focus on metadata accessibility and quality, each element of which can impact decision-making. To address issues like reading fatigue from metadata presentation, inefficient methods for results searching, filtering, and selection, along with other unresolved user-centric issues on current platforms. These prototypes concentrate on enhancing metadata-related features. They include a redesigned homepage, an improved search bar, better sorting options, an enhanced search result display, a metadata comparison tool, and a navigation guide. Our aim is to improve usability for a wide range of users, including both developers and researchers.

Human-Computer Interaction