10.2312/eurovisshort.20161164
Visualizing TimeDependent Data Using Dynamic tSNE
Rauber, Paulo E.
Falcão, Alexandre X.
Telea, Alexandru C.
The Eurographics Association
2016

9783038680147
http://diglib.eg.org/bitstream/handle/10.2312/eurovisshort20161164/073077.pdf
5 pages
Many interesting processes can be represented as timedependent datasets. We define a timedependent dataset as a sequence of datasets captured at particular time steps. In such a sequence, each dataset is composed of observations (highdimensional real vectors), and each observation has a corresponding observation across time steps. Dimensionality reduction provides a scalable alternative to create visualizations (projections) that enable insight into the structure of such datasets. However, applying dimensionality reduction independently for each dataset in a sequence may introduce unnecessary variability in the resulting sequence of projections, which makes tracking the evolution of the data significantly more challenging. We show that this issue affects tSNE, a widely used dimensionality reduction technique. In this context, we propose dynamic tSNE, an adaptation of tSNE that introduces a controllable tradeoff between temporal coherence and projection reliability. Our evaluation in two timedependent datasets shows that dynamic tSNE eliminates unnecessary temporal variability and encourages smooth changes between projections.
EuroVis 2016  Short Papers
Multidimensional and Geospatial Visualization
73
77
Paulo E. Rauber, Alexandre X. Falcão, and Alexandru C. Telea
Categories and Subject Descriptors (according to ACM CCS): Humancentered computing  Information visualization; Computing methodologies  Dimensionality reduction and manifold learning
Humancentered computing Information visualization Computing methodologies Dimensionality reduction and manifold learning
073077