Academia.eduAcademia.edu

A live demo

1987

This chapter contains sections titled: The game of “life”, Echoing, Tracing, How to breed gliders

Interactive and Dynamic Statistical Graphics - An Overview Jürgen Symanzik Utah State University, Logan, UT, USA *e-mail: [email protected] WWW: http://www.math.usu.edu/~symanzik Contents Data, Terms, Citations, and Definitions n Main Concepts n Graphical Software n Live Demo n Conclusion n Places Data n “Places” data set: – 329 cities in the U.S. – 9 measures of livability (early 1980’s): Climate & Terrain, Housing Cost, Health Care & Environment, Crime, Transportation, Education, The Arts, Recreation, and Economics. – Published in Places Rated Almanac (Boyer and Savageau, 1981), copyrighted by Rand McNally – Latitude and longitude added by Paul Tukey Terms Interactive & Dynamic Statistical Graphics (DSG) n Exploratory Data Analysis (EDA) n Exploratory Spatial Data Analysis (ESDA) n Visual Data Mining (VDM) n Visual Analysis/Visual Analytics (VA) n Data Mining (DM) n Citations n John W. Tukey (1977): EDA “is detective work - numerical detective work - or counting detective work - or graphical detective work.” n Edward J. Wegman (2000): “Data Mining is exploratory data analysis with little or no human interaction using computationally feasible techniques, i.e., the attempt to find interesting structure unknown a priori.” DSG/VDM (1) n Working Definition for DSG/VDM: – Find structure (cluster, unusual observations) in large and not necessarily homogeneous data sets based on human perception using graphical methods and user interaction – Goal or expected outcome of exploration usually unknown in advance DSG/VDM (2) n First uses of the term VDM: – Cox, Eick, Wills, Brachman (1997): Visual Data Mining: Recognizing Telephone Calling Fraud, Data Mining and Knowledge Discovery, 1:225-231. – Inselberg (1998): Visual Data Mining with Parallel Coordinates, Computational Statistics, 13(1):47-63. DSG Concepts (1) Scatterplots and Scatterplot Matrices n Brushing and Linked Brushing/Linked Views n Focusing, Zooming, Panning, Slicing, Rescaling, and Reformatting n Rotations and Projections n Grand Tour n Parallel Coordinate Plots n DSG Concepts (2) Projection Pursuit and Projection Pursuit Guided Tours n Pixel or Image Grand Tours n Andrews Plots n Density Plots, Binning, and Brushing with Hue and Saturation n Special DSG techniques for Categorical Data n Scatterplots and Linked Brushing XGobi Scatterplot Matrix and Density Plot ExplorN Parallel Coordinate Plots ExplorN Grand Tour – Continuous random sequence of projections from n dimensions into 2 (or more) dimensions. Graphical Software Origin: PRIM-9 n REGARD, MANET, and Mondrian Family n EXPLOR4, HyperVision, ExplorN, and CrystalVision Family n DataViewer, XGobi, and GGobi Family n Origin of DSG Software: PRIM-9 “Picturing, Rotation, Isolation and Masking in up to 9 Dimensions” n Initiated in the early 1970's by M. A. Fisherkeller, J. H. Friedman, and J. W. Tukey n Main features: n – Projections – Isolations and Masking DSG Software: REGARD, MANET, and Mondrian Initiated in the late 1980's by John Haslett and Antony Unwin at Trinity College, Dublin, Ireland n Continued by Antony Unwin and collaborators at University of Augsburg, Germany n Other main collaborators: Heike Hofmann, Martin Theus, Adalbert Wilhelm, and Graham Wills n REGARD “Radical Effective Graphical Analysis of Regional Data” n Early 1990’s, Macintosh n High interaction graphics tools for spatial data n Map window that is linked to statistical displays n MANET “Missings Are Now Equally Treated” n Mid/Late 1990’s, Macintosh n http://www1.math.uni-augsburg.de/Manet n Graphics for continuous and discrete data n Keeps track of missing values in graphics n Mondrian Early 2000’s, JAVA n http://www.rosuda.org/Mondrian/ n Visualization of categorical and geographic data n DSG Software: EXPLOR4, HyperVision, ExplorN, and CrystalVision Initiated in the late 1980's by Dan Carr and Ed Wegman at George Mason University n Other main collaborators: Qiang Luo and Wesley L. Nicholson n EXPLOR4 Late 1980’s, VAX 11/780, Fortran n Main features: n – Rotations – Scatterplots & Scatterplot Matrix – Stereoscopic Views HyperVision Late 1980’s, IBM RT & MS-DOS, Pascal n Main features: n – Real Time Rotations – 2D & 3D Scatterplots & Scatterplot Matrix – Parallel Coordinate Plots – Color Histograms ExplorN Mid 1990’s, SGI n ftp://www.galaxy.gmu.edu/pub/software/ n Interactive environment for exploring multivariate data: n – Advanced Parallel Coordinates Displays – 3D Surfaces – Stereoscopic Displays CrystalVision Early 2000’s, PCs n ftp://www.galaxy.gmu.edu/pub/software/ n Main features: n – Parallel coordinate plots – Scatterplots – Grand tour animations DSG Software: DataViewer, XGobi, and GGobi Initiated in the mid 1980's by Andreas Buja, Deborah F. Swayne, and Dianne Cook at the University of Washington, Bellcore, AT&T Bell Labs, and Iowa State University n Other main collaborators: Catherine Hurley, John A. McDonald, and Duncan Temple Lang n DataViewer Mid 1980’s, Symbolics Lisp Machine n Main features: n – Linked windows – Focusing – Projections such as 3D rotations and grand tour XGobi Early 1990’s through early 2000’s n UNIX and Linux platforms n http://www.research.att.com/areas/stat/xgobi/ n Main features: n – Linked views – Linked brushing – Univariate, bivariate, and multivariate views – Grand tour – Links to other software GGobi Early 2000’s n PCs, UNIX and Linux platforms n http://www.ggobi.org/ n Main features: n – Very similar to XGobi – Multiple plot windows – Uses GTK+ graphical toolkit Live Demo n GGobi n CrystalVision n Mondrian Places Data in VRGobi Places Data as Micromaps Conclusions Visual approach effective to see unexpected structure in data n Combination of different techniques most effective n Can be used for almost all types of data n Main Reference: Symanzik, J. (2004): Interactive and Dynamic Graphics, In: Gentle, J. E., Härdle, W., Mori, Y. (Eds.), Handbook of Computational Statistics Concepts and Methods, Springer, Berlin/Heidelberg, 293336. Questions ???