Structuring Visual Exploratory Analysis of Skill Demand
42 Pages Posted: 7 Jul 2018 Publication Status: Accepted
The analysis of increasingly large and diverse data for meaningful interpretation and question answering is handicapped by human cognitive limitations. Consequently, semi-automatic abstraction of complex data within structured information spaces becomes increasingly important, if its knowledge content is to support intuitive, exploratory discovery. Exploration of skill demand is an area where regularly updated, multi-dimensional data may be exploited to assess capability within the workforce to manage the demands of the modern, technology- and data-driven economy. The knowledge derived may be employed by skilled practitioners in defining career pathways, to identify where, when and how to update their skillsets in line with advancing technology and changing work demands. This same knowledge may also be used to identify the combination of skills essential in recruiting for new roles. To address the challenges inherent in exploring the complex, heterogeneous, dynamic data that feeds into such applications, we investigate the use of an ontology to guide structuring of the information space, to allow individuals and institutions to interactively explore and interpret the dynamic skill demand landscape for their specific needs. As a test case we consider the relatively new and highly dynamic field of Data Science, where insightful, exploratory data analysis and knowledge discovery are critical. We employ context-driven and task-centered scenarios to explore our research questions and guide iterative design, development and formative evaluation of our ontology-driven, visual exploratory discovery and analysis approach, to measure where it adds value to users' analytical activity. Our findings reinforce the potential in our approach, and point us to future paths to build on.
Keywords: domain modeling, knowledge discovery, visual exploration, ontology-guided visual analytics, trend identification, demand analysis
Suggested Citation: Suggested Citation