Analytic Component Tackling tables Analytic component Tasks and - PowerPoint PPT Presentation

Focus on Tables Exercise: Sketch 2 ways to visualize each table Information Visualization Dataset Types Spatial Net Tables Tables Networks Fields (Continuous) Geometry (Spatial) Attributes (columns) Furthest BPM T1 BPM T2 BPM T3 Age Best 100 m Sex Jump Items Link Grid of positions Amy 90 130 150 (rows) Amy 16 13.2 5.2 F Node (item) Cell Basil 70 110 109 Basil 18 12.4 4.2 F Tamara Munzner Tables Position Cell containing value Node Clara 60 140 141 Clara 14 14.1 2.5 F em) Department of Computer Science Attributes (columns) Multidimensional Table Trees Desmond 84 100 108 Desmond 22 10.01 6.3 M University of British Columbia Value in cell Charles 81 110 130 Charles 19 11.3 5.3 M Lect 6/7, 23/28 Jan 2020 Value in cell • socrative: answer when done https://www.cs.ubc.ca/~tmm/courses/436V-20 2 3 4 Analytic Component Tackling tables Analytic component Tasks and techniques Keys and values Age Gender Height Tables Bob 25 M 181 • homogeneity • key Alice 22 F 185 Magnitude Distribution Deviation Correlation Attributes (columns) Chris 19 M 175 –same data type? same scales? –independent attribute Items BPM 1 BPM 2 BPM 3 (rows) Ranking Part to whole Change over Time Bob 65 120 145 –used as unique index to look up items Alice 80 135 185 Cell containing value –simple tables: 1 key Chris 45 115 135 • need different approaches based on scale Multidimensional Scaling Scatterplot Matrices –multidimensional tables: multiple keys [Doerk 2011] Multidimensional Table [Bostock] –how many attributes? • value • up to ~50: tractable with direct visual encoding Pixel-based visualizations / • thousands: need transformations / analytical methods heat maps –dependent attribute, value of cell Value in cell –how many items? • classify arrangements by key count Parallel Coordinates • up to 1K: tractable with direct visual encoding –0, 1, 2, many... [Bostock] • >> 10K: need transformations / analytical methods [Chuang 2012] Express Values 1 Key 2 Keys 3 Keys Many Keys no / little analytics strong analytics List Matrix Volume Recursive Subdivision component https://github.com/ft-interactive/chart-doctor/tree/master/visual-vocabulary 
 5 6 https://gramener.github.io/visual-vocabulary-vega/#/Magnitude/ 7 8 0 Keys: Express values (magnitudes) Idiom: scatterplot Scatterplots: Encoding more channels Scatterplot tasks Express Values • express values • additional channels for point marks • correlation –quantitative attributes –color • no keys, only values –size (bubbleplots) Express Values 1 Key 2 Keys 3 Keys Many Keys List Matrix Volume Recursive Subdivision • square root since area grows quadratically, radius is misleading –data –shape • 2 quant attribs –mark: points • clusters/groups, and clusters vs classes https://www.mathsisfun.com/data/scatter-xy-plots.html –channels • horiz + vert position –tasks • find trends, outliers, distribution, correlation, clusters –scalability • hundreds of items https://observablehq.com/@d3/scatterplot-with-shapes https://www.d3-graph-gallery.com/graph/bubble_basic.html [A layered grammar of graphics. Wickham. Journ. Computational and Graphical Statistics 19:1 (2010), 3–28.] 9 10 11 https://www.cs.ubc.ca/labs/imager/tr/2014/DRVisTasks/ 12 Some keys Some keys: Categorical regions Idiom: bar chart Separated and Aligned but not Ordered 100 100 • one key, one value 75 75 Separate Order Align 50 50 –data 25 25 • 1 categ attrib, 1 quant attrib Express Values 1 Key 2 Keys 3 Keys Many Keys 0 0 List Matrix Volume Recursive Subdivision –mark: lines –channels Animal Type Animal Type • length to express quant value • regions : contiguous bounded areas distinct from each other • spatial regions: one per mark –using space to separate (proximity) – separated horizontally, aligned vertically –following expressiveness principle for categorical attributes – ordered by quant attrib • use ordered attribute to order and align regions » by label (alphabetical), by length attrib (data-driven) –task 1 Key 2 Keys 3 Keys Many Keys • compare, lookup values List Matrix Volume Recursive Subdivision –scalability LIMITATION: Hard to know rank. What’s the 4 th most? The 7 th ? • dozens to hundreds of levels for key attrib 13 14 15 [Slide courtesy of Ben Jones]