VISUALIZATION WORKSHOP Amit Chourasia Visualization Services Group San Diego Supercomputer Center, UCSD AGENDA 01:30pm – 02:15pm SESSION 1A: Information visualization (Lecture) Motivation for visualization Assay of Techniques and their Application 02:15pm – 03:30pm SESSION 1B: Scientific visualization (Lecture) Assay of Techniques Application Use Cases Best Practices 03:00pm – 03:30pm Q&A, BREAK 03:30pm – 04:45pm SESSION 2: Visualization with VisIt (Hands on) 04:45pm – 05:00pm Q & A, Adjourn Be Prepared for Session 2: Visualization Hands On Make sure you have VisIt (2.5.2*): http://visit.llnl.gov/executables.html *Preferred version Sample Data: Unzip sample data to your Desktop https://wci.llnl.gov/codes/visit/2.3.0/VisItClassData.zip Sphere Data http://users.sdsc.edu/~amit/forums/sphere.zip TUTORIAL GOALS Session 1 Visualization concepts Visualization use cases Best practices in visualization Strengths and limitation of visualization Introduction to VisIt software – Perform basic tasks in VisIt Session 2 Perform Basic and Intermediate tasks with VisIt Remote visualization with VisIt on Gordon SESSION 1A: INFORMATION VISUALIZATION VIZ MISCONCEPTIONS • I am not an artist thus can’t do viz • Viz is an art, not science or engineering • Viz is a one-time task • Viz is useful only for communication • Viz SU’s are insignificant Why should you care about visualization? HOW MANY 3’S IN FIRST 350 NUMBERS OF PI? 3.14159265358979323846264338327950288419716939937510582097494459230781640628620 8998628034825342117067982148086513282306647093844609550582231725359408128481117 450284102701938521105559644622948954930381964428810975665933446128475648233786 783165271201909145648566923460348610454326648213393607260249141273724587006606 3155881748815209209628292540917153643 HOW MANY 3’S IN FIRST 350 NUMBERS OF PI? 3.14159265358979323846264338327950288419716939937 8 510582097494459230781640628620899862803482534211 3 706798214808651328230664709384460955058223172535 5 940812848111745028410270193852110555964462294895 3 493038196442881097566593344612847564823378678316 4 527120190914564856692346034861045432664821339360 6 726024914127372458700660631558817488152092096282 2 92540917153643 2 Total 33 What did you observe? PREATTENTIVE PROCESSING Unconsciously gathering information from the environment Preattentive Attributes(partial list): position, orientation, scale color, brightness, saturation shape, texture WHAT IS VISUALIZATION? Working Definition Encoding/mapping data into a visual representation to gain/extract insights DATA • Text (ASCII or Binary) • Images (confocal microscopy, satellite imagery) • High Dimensional (structured and Unstructured) • Mesh discretizes space into points and cells -1D, 2D, 3D Slide: Courtesy of Sean Ahern, NICS VARIABLES • Scalars • Vectors • Tensors • Multi-dimensional Slide: Courtesy of Sean Ahern, NICS MOTIVATION FOR VISUALIZATION Create visual representations based on underlying data that are • Concise (Yes) • Unambiguous (Preferably) • Intuitive (Trainable) • Interactive (Desirable) • Scalable (We wish) VISUALIZATION BUILDING BLOCKS Viz Elements Viz Attributes View Attributes • Glyphs (symbols: e.g. Alphabets, Arrows, Points) • Transforms • Viewpoint (Position, • Lines • Projection Rotation, Scale) • Triangles (Orthographic, • Color Perspective) • Voxels* (volume element) *Cannot be directly represented on • Opacity • Canvas displays Viz Reinforcement • Texture • Light • Distortion (e.g. displacement) • Motion (e.g. Camera, time steps) • Filter (e.g. threshold, resample, subset, slice, clip) • Add Context (e.g. Connectivity, Map Overlay) VISUALIZATION TECHNIQUES Scientific Viz Information Viz • Color Map • Plots (scatter, bar, pie …) • Contours, Isosurface And Explicit • Heatmaps Geometry • Parallel Coordinates • Volumetric • Treemaps • Streamlines • Partition Maps • Line Integral Convolution • Flow Maps • Topological • Networks • Glyphs PLOTS AND CHARTS Image: d3js gallery by Michael Bostock I II III IV x1 y1 x2 y2 x3 y3 x4 y4 10 8.04 10 9.14 10 7.46 8 6.58 8 6.95 8 8.14 8 6.77 8 5.76 13 7.58 13 8.74 13 12.74 8 7.71 9 8.81 9 8.77 9 7.11 8 8.84 11 8.33 11 9.26 11 7.81 8 8.47 14 9.96 14 8.1 14 8.84 8 7.04 6 7.24 6 6.13 6 6.08 8 5.25 4 4.26 4 3.1 4 5.39 19 12.5 12 10.84 12 9.13 12 8.15 8 5.56 7 4.82 7 7.26 7 6.42 8 7.91 5 5.68 5 4.74 5 5.73 8 6.89 mean(X) = 9, variance(X) = 11 mean(Y) = 7.5, variance(Y) = 4.12 linear regression lineY = 3 + 0.5*X correlation(X,Y) = 0.816 Anscombe’s Quartet 12 12 10 10 8 8 Y2 Y1 6 6 4 4 2 2 0 0 0 2 4 6 8 10 12 14 16 0 2 4 6 8 10 12 14 16 X1 X2 14 14 12 12 10 10 8 8 Y3 Y4 6 6 4 4 2 2 0 0 0 2 4 6 8 10 12 14 16 0 5 10 15 20 X3 X4 Anscombe 1973,TheAmerican Statistician HORIZON CHARTS Image: horizon charts by Michael Bostock HEATMAPS Process: Map scalar data to a color table Visual Validation Utility: To investigate range of data Swift error diagnostic and visual validation Images: Mathworks.com (Heatmap example) HEATMAPS = COLORMAPS = PSUEDOCOLOR PLOTS PARALLEL COORDINATES Process: Connect data in a pair wise manner on Y-Y axes PARALLEL COORDINATE - INTERACTIVITY • Filter example • Hover example • All hit stats example • Sophisticated example Utility: Summarize high dimensional data Find trends and relationships PARALLEL SETS Image: Titanic Survivors by Michael Bostock TREEMAP Process: Recursive mapping of hierarchical data into rectangles Treemap designed by Ben Shneiderman, UMD (1990) History and examples Utility: Compare tree structures and attributes of varying depth Image: Treemap Demo, UMD TREEMAP VARIATIONS Circular Treemap Voronoi Treemap Image: Kai Wetzel Circa (2005): Michael Blazer and Deussen TREEMAP IN ACTION: MAP OF THE MARKET WSJ’s SmartMoney TREEMAP -INTERACTION Image: Treemap example by Michael Bostock PARTITION MAP Process: transform a hierarchical data into a linearly proportionate rectangles Interactive Link Utility: examine proportionate distribution of hierarchical data FLOW MAP Process: connect items depicting flow in proportionate manner Utility: identify relationship, nature or flow (swell and attrition) and corresponding events Image: Charles Joseph Minard (1869) FLOW DIAGRAMS Image: Google Analytics SANKEY DIAGRAMS Image: Sankey Diagram by Michael Bostock NETWORKS Process: topologically represent hierarchical data Utility: Show and investigate relationships Image: Force Directed Layout by Michael Bostock LONDON UNDERGROUND MAP 1932 Image: A History of the London Tube Maps LONDON UNDERGROUND MAP 1933 Image: London Transport website DECISION TREE Image: “512 Paths to the White House” by Mike Bostock and Shan Carter, New York Times HIERARCHICAL NETWORK PRESENTED BY FILL OUT YOUR BRACKET Second Round Third Round Regional Regional National N A TION A L BRACKET DA Y National Regional Regional Third Round Second Round MA RCH 15- 16 MARCH 17-18 Semifinals Finals Semifinals MARCH 12 Semifinals Finals Semifinals MA RCH 17- 18 MA RCH 15- 16 MA RCH 2 2- 23 MA RCH 24 - 25 MA RCH 31 MA RCH 31 MA RCH 24 - 25 MARCH 22- 23 First Four ® 16 Miss. Val. (21-12) 58 14 BYU (25-8) 78 First Round* 16 Lamar (23-11) 59 12 California (24-9) 54 Mar 13 Mar 13 DAYTON Mar 14 Mar 14 S 16 W . Ky. (15-18) 59 W 14 Iona (25-7) 72 MA RCH 13- 14 MW 16 Vermont (23-11) 71 MW 12 S. Fla. (20 -13) 65 W atch On 1 Kentucky (32-2) 81 1 Syracuse (31-2) 72 1 Kentucky 87 1 Syracuse 75 16 W estern Ky. 66 16 UNC Asheville (24-9) 65 Louisville Pittsburgh 1 Kentucky 102 1 Syracuse 64 Mar 17 Mar 17 8 Iowa St. (22-10) 77 8 Kansas St. (21-10) 70 8 Iowa St. 71 8 Kansas St. 59 9 Connecticut (20 -13) 64 9 Southern Miss. (25-8) 64 Atlanta 1 Kentucky 82 1 Syracuse 70 Boston Mar 23 Mar 22 5 W ichita St. (27-5) 59 5 Vanderbilt (24-10) 79 12 VCU 61 5 Vanderbilt 57 12 VCU (28-6) 12 Harvard (26-4) 62 Albuquerque 70 Portland 4 Indiana 90 4 W isconsin 63 Mar 17 Mar 17 4 Indiana (25-8) 79 4 W isconsin (24-9) 73 4 Indiana 63 4 W isconsin 60 13 New Mexico St. (26-9) 66 13 Montana (25-6) 49 SOUTH 1 Kentucky 69 2 Ohio St. 62 EAST 6 ® 6 Cincinnati (24-10) UNLV (26-8) 64 ATLANTA Mar 31 Final Four Mar 31 BOSTON 65 11 Colorado 63 March 25 March 24 6 Cincinnati 62 11 Colorado (23-11) NEW ORLEANS 11 Texas (20 -13) 68 Nashville 59 Albuquerque 3 Baylor 75 MARCH 31 AND APRIL 2 6 Cincinnati 66 Mar 17 Mar 18 3 Baylor (27-7) 68 3 Florida St. (24-9) 66 3 Baylor 80 3 Florida St. 56 14 S. Dakota St. (27-7) 60 14 St. Bonaventure (20 -11) 63 Atlanta 3 Baylor 70 2 Ohio St. 77 Boston 7 Notre Dame (22-11) 63 Mar 23 Mar 22 7 Gonzaga (25-6) 77 10 Xavier 70 National 7 Gonzaga 66 10 Xavier (21-12) 67 10 West Virginia (19-13) 54 Pittsburgh Greensboro 10 Xavier 70 Championship 2 Ohio St. 81 Mar 18 Mar 17 2 Duke (27-6) 70 A PRIL 2 2 Ohio St. (27-7) 78 15 Lehigh 58 2 Ohio St. 73 15 Lehigh (26-7) 75 15 Loyola (MD) (24-8) 59 1 Kentucky 67 Kentucky 2 Kansas 59 1 Michigan St. (27-7) 89 1 North Carolina (29-5) 77 1 Michigan St. 65 1 UNC 87 16 LIU Brooklyn (25-8) 67 16 Vermont 58 Columbus 1 Mich. St. 44 1 UNC 73 Greensboro Mar 18 Mar 18 8 Memphis (26-8) 54 8 Creighton (28-5) 58 9 Saint Louis 61 8 Creighton 73 9 Saint Louis (25-7) 61 9 Alabama (21-11) 57 Phoenix 4 Louisville 72 1 UNC 67 St. Louis 5 New Mexico (27-6) 75 Mar 22 Mar 23 5 Temple (24-7) 44 5 New Mexico 56 12 South Fla. 56 12 Long Beach St. (25-8) 68 12 South Fla. 58 Portland 4 Louisville 57 13 Ohio 65 Nashville Mar 17 Mar 18 4 Louisville (26-9) 69 4 Michigan (24-9) 60 4 Louisville 59 13 Ohio 62 13 Davidson (25-7) 62 13 Ohio (27-7) 65 WEST 4 Louisville 61 2 Kansas 64 MIDWEST 6 Murray St.