New insights into the suitability of the third dimension for visualizing multivariate/multidimensional data: A study based on loss of quality quantification
Restricted accessResearch articleFirst published online January, 2016
New insights into the suitability of the third dimension for visualizing multivariate/multidimensional data: A study based on loss of quality quantification
Most visualization techniques have traditionally used two-dimensional, instead of three-dimensional representations to visualize multidimensional and multivariate data. In this article, a way to demonstrate the underlying superiority of three-dimensional, with respect to two-dimensional, representation is proposed. Specifically, it is based on the inevitable quality degradation produced when reducing the data dimensionality. The problem is tackled from two different approaches: a visual and an analytical approach. First, a set of statistical tests (point classification, distance perception, and outlier identification) using the two-dimensional and three-dimensional visualization are carried out on a group of 40 users. The results indicate that there is an improvement in the accuracy introduced by the inclusion of a third dimension; however, these results do not allow to obtain definitive conclusions on the superiority of three-dimensional representation. Therefore, in order to draw further conclusions, a deeper study based on an analytical approach is proposed. The aim is to quantify the real loss of quality produced when the data are visualized in two-dimensional and three-dimensional spaces, in relation to the original data dimensionality, to analyze the difference between them. To achieve this, a recently proposed methodology is used. The results obtained by the analytical approach reported that the loss of quality reaches significantly high values only when switching from three-dimensional to two-dimensional representation. The considerable quality degradation suffered in the two-dimensional visualization strongly suggests the suitability of the third dimension to visualize data.
HoffmanPEGrinsteinGG. A survey of visualizations for high-dimensional data mining. In: FayyadUGrinsteinGGWierseA (eds) Information visualization in data mining and knowledge discovery. San Francisco, CA: Morgan Kaufmann Publishers Inc., 2002, pp. 47–82.
2.
WongPCBergeronRD. 30 years of multidimensional multivariate visualization. In: Scientific Visualization, Overviews, Methodologies, and Techniques (eds NielsonGMHagenHMüllerH), pp. 3–33. Washington, DC, USA: IEEE Computer Society.
3.
ToryMMöllerT. A model-based visualization taxonomy. Technical Report CMPT-TR2002-06, 2002. Computing Science Department, Simon Fraser University.
4.
KeimDA. Designing pixel-oriented visualization techniques: theory and applications. IEEE T Vis Comput Gr2000; 6(1): 59–78.
5.
KeimDA. Information visualization and visual data mining. IEEE T Vis Comput Gr2002; 8(1): 1–8.
6.
KeimDAKriegelH-P. Visualization techniques for mining large databases: a comparison (special issue on data mining). IEEE T Knowl Data En1996; 8(6): 923–938.
7.
HartiganJA. Printer graphics for clustering. J Stat Comput Sim1975; 4(3): 187–213.
8.
FurnasGWBujaA. Prosection views: dimensional inference through sections and projections. J Comput Graph Stat1994; 3(4): 323–385.
9.
TweedieLSpenceR. The Prosection Matrix: a tool to support the interactive exploration of statistical models and data. Computation Stat1998; 13: 65–76.
10.
Van WijkJJVan LiereR. HyperSlice: visualization of scalar functions of many variables. In: Proceedings of the 4th conference on Visualization ’93 (VIS ’93) (ed Bergeron DDNielsonG), October 1993, pp. 119–125. Washington, DC, USA:IEEE Computer Society.
11.
WongPCCrabbAHBergeronRD. Dual multiresolution HyperSlice for multivariate data visualization. In: Proceedings of the 1996 IEEE Symposium on Information Visualization (INFOVIS ’96) (ed CardSKEickSGGershonND), San Francisco, CA, 28–29 October 1996, pp. 74–75.
12.
AlpernBCarterL. The hyperbox. In: Proceedings of the 2nd conference on Visualization ’91 (VIS ’91), (ed NielsonGMRosenblumL), 22–25 October 1991, pp. 133–139. Los Alamitos, CA, USA: IEEE Computer Society Press.
13.
InselbergA. Multidimensional detective. In: Proceedings of the 1997 IEEE symposium on information visualization (INFOVIS ’97), 20–21 October 1997, pp. 100–107. Washington, DC: IEEE Computer Society.
14.
InselbergA. The plane with parallel coordinates. Visual Comput1985; 1(2): 69–91.
15.
AndrewsDF. Plots of high-dimensional data. Biometrics1972; 28(1): 125–136.
16.
HoffmanPE. Table visualizations: a formal model and its applications. PhD Thesis, Computer Science Department, University of Massachusetts Lowell, Lowell, MA, 1999.
17.
BeddowJ. Shape coding of multidimensional data on a microcomputer display. In: Proceedings of the 1st conference on Visualization ’90 (VIS ’90), (ed KaufmanA), 23–26October1990, pp. 238–246. Los Alamitos, CA, USA:IEEE Computer Society Press.
18.
RaoRCardSK. The Table Lens: merging graphical and symbolic representations in an interactive focus + context visualization for tabular information. In: Proceedings of the SIGCHI Conference on Human Factors in Computing Systems (CHI ’94), (eds AdelsonBDumaisSOlsonJ ), 1994, pp. 318–322. New York, NY, USA: ACM.
19.
PeanoG. Sur une courbe, qui remplit toute une aire plane. Math Ann1890; 36: 157–160.
20.
HilbertD. Ueber die stetige Abbildung einer Line auf ein Flächenstück. Math Ann1891; 38(3): 459–460.
21.
MortonGM. A computer oriented geodetic data base and a new technique in file sequencing. Technical Report, IBM. 1966. Ottawa, Canada.
22.
KeimDAAnkerstMKriegelH-P. Recursive pattern: a technique for visualizing very large amounts of data. In: Proceedings of the 6th conference on visualization ’95 (VIS ’95), Atlanta, GA, 29 October–3 November 1995, pp. 279–286. Washington, DC: IEEE Computer Society.
AnkerstMKeimDAKriegelH-P. Circle segments: a technique for visually exploring large multidimensional data sets. In: Visualization ’96, hot topic session, San Francisco, CA, 27October–1November1996.
25.
KeimDAHaoMDayalUHsueM. Pixel bar charts: a visualization technique for very large multi-attribute data sets. Information Visualization2002; 1(1): 20–34.
26.
MihalisinTGawlinskiETimlinJSchweglerJ. Visualizing a scalar field on an N-dimensional lattice. In: Proceedings of the 1st conference on Visualization ’90 (VIS ’90) (ed KaufmanA), San Francisco, CA, 23–26 October 1990, pp. 255–262. Los Alamitos, CA, USA: IEEE Computer Society Press.
27.
MihalisinTTimlinJSchweglerJ. Visualization and analysis of multi-variate data: a technique for all fields. In: Proceedings of the 2nd conference on visualization ’91 (VIS ’91), San Diego, CA, 22–25 October 1991, pp 171–178. Los Alamitos, CA, USA: IEEE Computer Society Press.
LeBlancJWardMOWittelsN. Exploring N-dimensional databases. In: Proceedings of the 1st conference on visualization ’90 (VIS ’90) (ed KaufmanA), San Francisco, CA, 23-26 October 1990, pp. 230–237. Los Alamitos, CA, USA: IEEE Computer Society Press.
ShneidermanB. Tree visualization with tree-maps: 2-d space-filling approach. ACM T Graphic1992; 11(1): 92–99.
32.
De OliveiraMCFLevkowitzH. From visual data exploration to visual data mining: a survey. IEEE T Vis Comput Gr2003; 9(3): 378–394.
33.
ChernoffH. The use of faces to represent points in k-dimensional space graphically. J Am Stat Assoc1973; 68(342): 361–368.
34.
ChambersJMClevelandWSKleinerB. Graphical methods for data analysis (The Wadsworth statistics/probability series). Boston, MA: Duxbury Press, 1983.
35.
PickettRMGrinsteinGG. Iconographic displays for visualizing multidimensional data. IEEE Sys Man Cybern1988; 1: 514–519.
36.
LevkowitzH. Color icons: merging color and texture perception for integrated visualization of multiple parameters. In: Proceedings of the 2nd conference on Visualization ’91 (VIS ’91) (eds NielsonGMRosenblumL), San Diego, CA, 22–25 October, 1991, pp. 164–170. Los Alamitos, CA, USA: IEEE Computer Society Press.
37.
EfrosAAFreemanWT. Image quilting for texture synthesis and transfer. In: Proceedings of the 28th annual conference on computer graphics and interactive techniques (SIGGRAPH ’01), 2001, pp. 341–346. New York: ACM.
38.
WeiL-YLevoyM. Fast texture synthesis using tree-structured vector quantization. In: Proceedings of the 27th annual conference on Computer graphics and interactive techniques (SIGGRAPH ’00), pp. 479–488. New York, NY, USA: ACM Press/Addison-Wesley Publishing Co.
39.
WareCKnightW. Using visual texture for information display. ACM T Graphic1995; 14(1): 3–20.
40.
LeeJAVerleysenM. Nonlinear dimensionality reduction. New York; London: Springer, 2007.
41.
KeimDAAndrienkoGFeketeJ-D. Visual analytics: definition, process, and challenges. In:KerrenAStaskoJTFeketeJ-D. (eds) Information visualization. Vol. 4950. Berlin, Heidelberg: Springer-Verlag, 2008, pp. 154–175.
42.
WangJ. Geometric structure of high-dimensional data and dimensionality reduction. Beijing, China: Higher Education Press and Springer, 2012.
43.
St. JohnMFCowenMBSmallmanHS. The use of 2D and 3D displays for shape-understanding versus relative-position tasks. Hum Factors2001; 43(1): 79–98.
44.
WareC. Visual thinking: for design (Morgan Kaufmann series in interactive technologies). Elsevier Science, 2010.
45.
GrossmanTWigdorDBalakrishnanR. Exploring and reducing the effects of orientation on text readability in volumetric displays. In: Proceedings of the SIGCHI Conference on Human Factors in Computing Systems (CHI ’07). (ed RossonMBGilmoreDJ), San Jose, CA, 28 April–3 May 2007, pp. 483–492. New York, USA: ACM.
46.
GraciaAGonzálezSRoblesV. A methodology to compare dimensionality reduction algorithms in terms of loss of quality. Inform Sciences2014; 270: 1–27.
47.
Van OrdenKFBroylesJW. Visuospatial task performance as a function of two and three-dimensional display presentation techniques. Displays2000; 21(1): 17–24.
48.
ParkSHWoldstadJC. Multiple two-dimensional displays as an alternative to three-dimensional displays in telerobotic tasks. Hum Factors2000; 42(4): 592–603.
49.
ToryMKirkpatrickAEStella AtkinsM. Visualization task performance with 2D, 3D, and combination displays. IEEE T Vis Comput Gr2006; 12(1): 2–13.
50.
ToryMSpragueDWWuF. Spatialization design: comparing points and landscapes. IEEE T Vis Comput Gr2007; 13(6): 1262–1269.
51.
ToryMSwindellsCDreezerR. Comparing dot and landscape spatializations for visual memory differences. IEEE T Vis Comput Gr2009; 15(6): 1033–1040.
52.
SmallmanHSSt. JohnMFOonkHM. Information availability in 2D and 3D displays. IEEE Comput Graph2001; 21(5): 51–57.
53.
FabrikantSI. Spatial metaphors for browsing large data archives. PhD Thesis, University of Colorado Boulder, Boulder, CO, USA, 2000.
54.
WickensCD. Virtual reality and education. In: Proc. of the IEEE Int. Conf. on Systems, Man and Cybernetics, New York, 1992, vol 1, pp. 842–847.
55.
WickensCD. Virtual reality and education. IEEE Sys Man Cybern1992; 1: 842–847.
56.
WickensCDMerwinDHLinEL. Implications of graphics enhancements for the visualization of scientific data: dimensional integrality, stereopsis, motion, and mesh. Hum Factors1994; 36(1): 44–61.
57.
SedlmairMMunznerTToryM. Empirical guidance on scatterplot and dimension reduction technique choices. IEEE T Vis Comput Gr2013; 19(12): 2634–2643.
58.
GolubTRSlonimDKTamayoP. Molecular classification of cancer: class discovery and class prediction by gene expression monitoring. Science1999; 286(5439): 531–537.
59.
DudoitSFridlyandJSpeedTP. Comparison of discrimination methods for the classification of tumors using gene expression data. J Am Stat Assoc2002; 97(457): 77–87.
MukherjeeSTamayoPRogersS. Estimating dataset size requirements for classifying DNA microarray data. J Comput Biol2003; 10(2): 119–142.
62.
JolliffeIT. Principal component analysis. Berlin, Heidelberg: Springer-Verlag, 1986.
63.
HotellingH. Analysis of a complex of statistical variables into principal components. J Educ Psychol1933; 24(6): 417–441.
64.
GoldstoneW. Unity 3.x game development essentials (Community experience distilled). Birmingham: Packt Publishing, 2011.
65.
LorenzoC-M. A review of interoperability and possibilities for data analysis from virtual world environments. Int J Metadata Semant Ontol2011; 6(3/4): 234–240.
66.
KimMJWangXLovePED. Virtual reality for the built environment: a critical review of recent advances. J Inf Technol Constr2013; 18: 279–305.
67.
ManferdiniAMRemondinoF. A review of reality-based 3D model generation, segmentation and web-based visualization methods. Int J Herit Digit Era2012; 1(1): 103–124.
68.
HelbigCBauerH-SRinkK. Concept and workflow for 3d visualization of atmospheric data in a virtual reality environment for analytical approaches. Environ Earth Sci2014; 72(10): 3767–3780.
69.
RinkKBilkeLKolditzO. Visualisation strategies for environmental modelling data. Environ Earth Sci2014; 72(10): 3857–3868.
70.
CiocADjorgovskiSGDonalekC. Data visualization using immersive virtual reality tools. In: American Astronomical Society meeting abstracts, Long Beach Convention and Entertainment Center, Long Beach, CA, vol. 221, 6–10 January 2013. AAS.
71.
WestlingCEIBarkerLHealyA. Spatial data correlation: an interactive 3D visualisation tool for correlating the motion capture data streams from difference devices. In: Electronic visualisation and the arts (EVA ’13), London 29–30July2013. London: British Computing Society.
SammonJW. A nonlinear mapping for data structure analysis. IEEE T Comput1969; C-18(5): 401–409.
81.
DemartinesPHeraultJ. Curvilinear component analysis: a self-organizing neural network for nonlinear mapping of data sets. IEEE T Neural Networ1997; 8(1): 148–154.
SchölkopfBSmolaAMüllerK-R. Nonlinear component analysis as a kernel eigenvalue problem. Neural Comput1998; 10(5): 1299–1319.
84.
SchölkopfBSmolaAMüllerK-R. Kernel principal component analysis. Cambridge, MA: MIT Press, 1999, pp. 327–352.
85.
TenenbaumJBSilvaVLangfordJC. A global geometric framework for nonlinear dimensionality reduction. Science2000; 290(5500): 2319–2323.
86.
De SilvaVTenenbaumJB. Global versus local methods in nonlinear dimensionality reduction. Adv Neur In2003; 15: 705–712.
87.
TenenbaumJB. Matlab package for Isomap. Stanford, CA: Stanford University, 2000. Available at : http://isomap.stanford.edu/ (accessed 10 December 2013).
88.
RoweisSTSaulLK. Nonlinear dimensionality reduction by locally linear embedding. Science2000; 290(5500): 2323–2326.
89.
SaulLKRoweisST. Think globally, fit locally: unsupervised learning of low dimensional manifolds. J Mach Learn Res2003; 4: 119–155.
90.
DudaROHartPEStorkDG. Pattern classification and scene analysis. New York: Wiley, 2001.
91.
HastieTJTibshiraniRJFriedmanJJH. The elements of statistical learning: data mining, inference, and prediction (Springer series in statistics series). Darinka Springer & Janez, 2001.
BelkinMNiyogiP. Laplacian eigenmaps and spectral techniques for embedding and clustering. Adv Neur In2001; 14: 585–591.
94.
BelkinMNiyogiP. Laplacian eigenmaps for dimensionality reduction and data representation. Neural Comput2003; 15(6): 1373–1396.
95.
WeinbergerKQShaFSaulLK. Learning a kernel matrix for nonlinear dimensionality reduction. In: Proceedings of the twenty-first international conference on Machine learning (ICML ’04) (ed BanffAB), Canada, 4–8 July 2004, pp. 106. New York, NY, USA: ACM.
96.
WeinbergerKQSaulLK. Unsupervised learning of image manifolds by semidefinite programming. Int J Comput Vision2006; 70(1): 77–90.
97.
WeinbergerKQSaulLK. An introduction to nonlinear dimensionality reduction by maximum variance unfolding. In: Proceedings of the 21st national conference on artificial intelligence (AAAI ’06), July 16–20, 2006, Boston, Massachusettsvol. 2, pp. 1683–1686. AAAI Press
NadlerBLafonSCoifmanRR. Diffusion maps, spectral clustering and reaction coordinates of dynamical systems. Appl Comput Harmon A2006; 21(1): 113–127.
100.
LafonSLeeA B. Diffusion maps and coarse-graining: a unified framework for dimensionality reduction, graph partitioning, and data set parameterization. IEEE T Pattern Anal2006; 28(9): 1393–1403.
101.
Van der MaatenLHintonG. Visualizing high-dimensional data using t-SNE. J Mach Learn Res2008; 9: 2579–2605.
102.
VennaJKaskiS. Local multidimensional scaling. Neural Networks2006; 19(6–7): 889–899.
103.
ChenL. Local multidimensional scaling for nonlinear dimension reduction, graph layout and proximity analysis, PhD Thesis, University of Pennsylvania, Philadelphia, PA, 2006.
104.
ChenLBujaA. Local multidimensional scaling for nonlinear dimension reduction, graph drawing, and proximity analysis. J Am Stat Assoc2009; 104(485): 209–219.
LeeJAVerleysenM. Rank-based quality assessment of nonlinear dimensionality reduction. In: Proceedings of the European symposium on artificial neural network (ESANN ’08), Bruges, 23–25 April 2008, pp. 49–54. d-side publi.
107.
GoldbergYRitovY. Local procrustes for manifold embedding: a measure of embedding quality and embedding algorithms. Mach Learn2009; 77(1): 1–25.
108.
LeeJAVerleysenM. Quality assessment of nonlinear dimensionality reduction based on K-ary neighborhoods. J Mach Learn Res2008; 4: 21–35.
109.
MengDLeungYXuZ. A new quality assessment criterion for nonlinear dimensionality reduction. Neurocomputing2011; 74(6): 941–948.
110.
ZhangPYuanyuanRZhangB. A new embedding quality assessment method for manifold learning. Neurocomputing2012; 97: 251–266.
111.
MangasarianOLNick StreetWWolbergWH. Breast cancer diagnosis and prognosis via linear programming. Oper Res1995; 43: 570–577.
112.
GormanRPSejnowskiTJ. Analysis of hidden units in a layered network trained to classify sonar targets. Neural Networks1988; 1(1): 75–89.
Nick StreetWMangasarianOLWolbergWH. An inductive learning approach to prognostic prediction. In: Proceedings of the twelfth international conference on machine learning (ICML ’95), Lake Tahoe, CA, March 1995, pp. 522–530. San Mateo, CA: Morgan Kaufmann.
115.
WolbergWHNick StreetWHeiseyDM. Computerized breast cancer diagnosis and prognosis from fine-needle aspirates. Arch Surg1995; 130(5): 511–516.
116.
SetlurSRMertzKDHoshidaY. Estrogen-dependent signaling in a molecularly distinct subclass of aggressive prostate cancer. J Natl Cancer I2008; 100(11): 815–825.
117.
EvettIWSpiehlerEJ. Rule induction in forensic science. In: DuffinPH (ed.) Knowledge based systems. New York: Halsted Press, 1988, pp. 152–160.
118.
LittleMAMcsharryPERobertsSJ. Exploiting nonlinear recurrence and fractal scaling properties for voice disorder detection. Biomed Eng Online2007; 6(1): 23.
119.
LittleMAMcSharryPEHunterEJ. Suitability of dysphonia measurements for telemonitoring of Parkinson’s disease. IEEE T Biomed Eng2009; 56(4): 1015–1022.
120.
ShippMARossKNTamayoP. Diffuse large B-cell lymphoma outcome prediction by gene-expression profiling and supervised machine learning. Nat Med2002; 8(1): 68–74.
121.
DeFelipeJLopez-CruzPLBenavides-PiccioneR. New insights into the classification and nomenclature of cortical GABAergic interneurons. Nat Rev Neurosci2013; 14(3): 202–216.
122.
KhanJWeiJSRingnerM. Classification and diagnostic prediction of cancers using gene expression profiling and artificial neural networks. Nat Med2001; 7(6): 673–679.
123.
DiasDBMadeoRCBRochaT. Hand movement recognition for Brazilian sign language: a study using distance-based neural networks. In: Proceedings of the 2009 international joint conference on neural networks (IJCNN ’09), Atlanta, GA, 14–19 June 2009, pp. 2355–2362. New York: IEEE.
