Profile
|
Daniel Bündgens, Dipl.-Math., MBA |
Publications
CAVIR: Correspondence Analysis in Virtual Reality. Ways to a Valid Interpretation of Correspondence Analytical Point Clouds in Virtual Environments
Correspondence Analysis (CA) is frequently used to interpret correlations between categorical variables in the area of market research. To do so, coherences of variables are converted to a three-dimensional point cloud and plotted as three different 2D-mappings. The major challenge is to correctly interpret these plottings. Due to a missing axis, distances can easily be under- or overestimated. This can lead to a misclustering and misinterpretation of data and thus to faulty conclusions. To address this problem we present CAVIR, an approach for CA in Virtual Reality. It supports users with a virtual three-dimensional representation of the point cloud and different options to show additional information, to measure Euclidean distances, and to cluster points. Besides, the free rotation of the entire point cloud enables the CA user to always have a correct view of the data.
@Article{Graff2012,
Title = {{CAVIR}: {C}orrespondence {A}nalysis in {V}irtual {R}eality. {W}ays to a {V}alid {I}nterpretation of {C}orrespondence {A}nalytical {P}oint {C}louds in {V}irtual {E}nvironments},
Author = {Frederik Graff and Andrea B\"{o}nsch and Daniel B\"{u}ndgens and Torsten Kuhlen},
Journal = {{C}onference {P}roceedings: {I}nternational {M}asaryk {C}onference for {P}h.{D}. {S}tudents and {Y}oung {R}esearchers},
Year = {2012},
Pages = {653-662},
Volume = {3},
Url = {http://www.vedeckekonference.cz/library/proceedings/mmk_2012.pdf}
}
CAVIR: Correspondence Analysis in Virtual Reality
Correspondence Analysis (CA) is used to interpret correlations between categorical variables in the areas of social science and market research. To do so, coherences of variables are converted to a three-dimensional point cloud and plotted as several different 2D-mappings, each containing two axes. The major challenge is to correctly interpret these plottings. Due to a missing axis, distances can easily be under- or overestimated. This can lead to a misinterpretation and thus a misclustering of data.
To address this problem we present CAVIR, an approach for CA in Virtual Reality. It supports users with a three-dimensional representation of the point cloud and different options to show additional information, to measure Euclidean distances, and to cluster points. Besides, the motion parallax and a free rotation of the entire point cloud enable the CA expert to always have a correct view of the data.
Best Presentation Award!
@Article{Boensch2012,
Title = {{CAVIR}: {C}orrespondence {A}nalysis in {V}irtual {R}eality},
Author = {Andrea B\"{o}nsch and Frederik Graff and Daniel B\"{u}ndgens and Torsten Kuhlen},
Journal = {{V}irtuelle und {E}rweiterte {R}ealit\"at, 9. {W}orkshop der {GI}-{F}achgruppe {VR}/{AR}},
Year = {2012},
Pages = {49-60},
ISSN = {978-3-8440-1309-2}
Publisher = {Shaker Verlag},
}
