RadViz plots are commonly used to represent multidimensional data because they use the familiar notion of 2D points for encoding data elements, displaying the original data dimensions that act as springs for setting the x and y coordinates. However, this intuitive approach implies several drawbacks and produces misleading visualizations that can confuse the user, even while analyzing a single data point. The paper attacks this problem following the well known idea of changing the order of the dimensions and introducing ancillary visualizations to mitigate some of RadViz drawbacks. In particular, the paper defines the notion of point optimal disposition of the dimensions for a single data point, generalizes this concept to a set of data points, and proposes effective heuristics for dealing with the intractable problem of exploring all the (n-1)2 dispositions of the dimensions along the RadViz circumference. Additional views, visual quality metrics, and a circular grid superimposed on the RadViz complement the attribute reordering strategy and provide a better understanding of the actual plot of the data elements.

Towards Enhancing RadViz Analysis and Interpretation / Angelini, M.; Blasilli, G.; Lenti, S.; Palleschi, A.; Santucci, G.. - (2019), pp. 226-230. (Intervento presentato al convegno 2019 IEEE Visualization Conference, VIS 2019 tenutosi a Vancouver; Canada) [10.1109/VISUAL.2019.8933775].

Towards Enhancing RadViz Analysis and Interpretation

Angelini M.
;
Blasilli G.
;
Lenti S.
;
Palleschi A.
;
Santucci G.
2019

Abstract

RadViz plots are commonly used to represent multidimensional data because they use the familiar notion of 2D points for encoding data elements, displaying the original data dimensions that act as springs for setting the x and y coordinates. However, this intuitive approach implies several drawbacks and produces misleading visualizations that can confuse the user, even while analyzing a single data point. The paper attacks this problem following the well known idea of changing the order of the dimensions and introducing ancillary visualizations to mitigate some of RadViz drawbacks. In particular, the paper defines the notion of point optimal disposition of the dimensions for a single data point, generalizes this concept to a set of data points, and proposes effective heuristics for dealing with the intractable problem of exploring all the (n-1)2 dispositions of the dimensions along the RadViz circumference. Additional views, visual quality metrics, and a circular grid superimposed on the RadViz complement the attribute reordering strategy and provide a better understanding of the actual plot of the data elements.
2019
2019 IEEE Visualization Conference, VIS 2019
dimension arrangement; multiple views; RadViz
04 Pubblicazione in atti di convegno::04b Atto di convegno in volume
Towards Enhancing RadViz Analysis and Interpretation / Angelini, M.; Blasilli, G.; Lenti, S.; Palleschi, A.; Santucci, G.. - (2019), pp. 226-230. (Intervento presentato al convegno 2019 IEEE Visualization Conference, VIS 2019 tenutosi a Vancouver; Canada) [10.1109/VISUAL.2019.8933775].
File allegati a questo prodotto
File Dimensione Formato  
Angelini_Towards-Enhancing_2019.pdf

solo gestori archivio

Tipologia: Versione editoriale (versione pubblicata con il layout dell'editore)
Licenza: Tutti i diritti riservati (All rights reserved)
Dimensione 5.94 MB
Formato Adobe PDF
5.94 MB Adobe PDF   Contatta l'autore

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/11573/1387815
Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus 9
  • ???jsp.display-item.citation.isi??? 8
social impact