Fast interactive design of scatterplots for large data set visualisation

Simo Santala*

*Corresponding author for this work

Research output: Contribution to conferenceAbstractScientificpeer-review

Abstract

Scatterplots are one of the most common data visualisation methods. Designing scatterplots for large data sets is challenging, as overlapping markers are likely to cause loss of information. Poorly chosen marker opacity, shape, or size may lead to e.g. overplotting or diminished visibility of outliers. The challenge is amplified by having to wait for rendering every time the design is changed. To reduce designer effort, optimisation-based approaches to scatterplot design have been proposed, most comprehensive being an algorithm by Micallef et al. (2017) that applies image-based perceptual quality measures to automatic evaluation of scatterplots. However, their approach suffers also from poor rendering performance, discouraging usage in interactive applications. This paper presents an algorithm applying abstract rendering for efficiently updating scatterplot markers regardless of data set size. We show how our approach enables fast, interactive design and adjustment of overlap in scatterplots, demonstrated with a proof-of-concept visualisation tool.

Original languageEnglish
Number of pages6
DOIs
Publication statusPublished - 25 Apr 2020
MoE publication typeNot Eligible
EventACM SIGCHI Annual Conference on Human Factors in Computing Systems - Honolulu, United States
Duration: 26 Apr 202030 Apr 2020
https://chi2020.acm.org/

Conference

ConferenceACM SIGCHI Annual Conference on Human Factors in Computing Systems
Abbreviated titleACM CHI
Country/TerritoryUnited States
CityHonolulu
Period26/04/202030/04/2020
Internet address

Keywords

  • Algorithms
  • Data visualisation
  • Optimisation

Fingerprint

Dive into the research topics of 'Fast interactive design of scatterplots for large data set visualisation'. Together they form a unique fingerprint.

Cite this