TimeElide: Visual Analysis of Non-Contiguous Time Series Slices



Michael Oppermann, Luce Liu, Tamara Munzner

 External link (DOI) 

 View presentation:2021-10-28T15:00:00ZGMT-0600Change your timezone on the schedule page
2021-10-28T15:00:00Z

Exemplar figure, described by caption below — Screenshot of TimeElide: a domain-agnostic, open source visualization tool for the visual analysis of non-contiguous time series slices. In the left sidebar, users can choose between manual and automatic slicing, and select a visual encoding. The main view shows our proposed visualization technique, sparkboxes, which supports the analysis of within-slice and across-slice patterns. The presented visualization shows the number of per-player events during 22 soccer matches.

Fast forward

Direct link to video on YouTube: https://youtu.be/l70SBo4uiks

Keywords

Domain Agnostic, Data Abstractions & Types, Software Prototype, Taxonomy, Models, Frameworks, Theory, Temporal Data

Abstract

We introduce the design and implementation of TimeElide, a visual analysis tool for the novel data abstraction of non-contiguous time series slices, namely time intervals that contain a sequence of time-value pairs but are not adjacent to each other. This abstraction is relevant for analysis tasks where time periods of interest are known in advance or inferred from the data, rather than discovered through open-ended visual exploration. We present a visual encoding design space as an underpinning of TimeElide, and the new sparkbox technique for visualizing fine and coarse grained temporal structures within one view. Datasets from different domains and with varying characteristics guided the development and their analysis provides preliminary evidence of TimeElide's utility.