MS Pattern Explorer: Interactive Visual Exploration of Temporal Activity Patterns

Keywords

Temporal Data, Task Abstractions & Application Domains, Application Motivated Visualization, Health Informatics, Mixed Initiative Human-Machine Analysis, Multiple Sclerosis

Abstract

Novel digital data sources have created new opportunities for digital health research, allowing for detailed insight into individuals’ daily lives. In the case of multiple sclerosis (MS), a neurodegenerative condition presenting as a highly variable corpus of symptoms, sensing devices afford researchers novel methodology for non-invasive, in-the-wild tracking of individuals’ physical activity. However, high-resolution sensors produce data of high dimensionality, complicating the contextualization of MS activity patterns. MS Pattern Explorer is a tool developed iteratively through close collaboration with clinicians, and facilitates the exploration and search for activity patterns in the context of explanatory metadata attributes, both within- and between persons with MS. Our validation includes feedback provided by health experts, along with two case studies conducted by a clinical collaborator. Results show that MS Pattern Explorer can effectively perform within- and between-patient analyses in novel ways, including pattern exploration, search, and contextualization.