TY  - CONF
TI  - Aesthetics and Ordering in Stacked Area Charts
AU  - Mathiesen, Steffen Strunge
AU  - Schulz, Hans-Jörg
A2  - Basu, Amrita
A2  - Stapleton, Gem
A2  - Linker, Sven
A2  - Legg, Catherine
A2  - Manalo, Emmanuel
A2  - Viana, Petrucio
T3  - Lecture Notes in Computer Science
AB  - Stacked area charts are a common visualisation type for sets of time series. Yet, they are also known to be challenging to read, in particular if the time series exhibit much fluctuation or even abrupt changes. In this paper, we introduce a novel approach to improving the layout of stacked area charts by means of reordering the time series in the stack. This approach breaks down into two parts: First, we gather aesthetic criteria and define associated quality metrics for stacked area charts. Second, we use these quality metrics together with a new algorithm called UpwardsOpt to find orderings of the stacked time series that optimise a chart's aesthetic properties. The produced orderings guarantee optimality in the sense that no better result can be obtained by moving any individual time series to a different position in the stack. In use cases, we illustrate how the flexibility provided by the introduced aesthetic criteria can make data properties better visible, as well as how our ordering algorithm is able to yield better orderings than the state-of-the-art approach. An open source implementation of our algorithm is provided to facilitate its reuse and the reproducibility of our results.
C3  - Proceedings of the International Conference on Theory and Application of Diagrams (Diagrams'21)
DA  - 2021///
PY  - 2021
DO  - 10.1007/978-3-030-86062-2_1
VL  - 12909
SP  - 3
EP  - 19
PB  - Springer
SN  - 978-3-030-86061-5
ER  -