Lecture 7/8: block feedback: many people not seeing value of - PowerPoint PPT Presentation

Viz theory How to handle complexity: 4 families of strategies Scenario Lecture 7/8: • block feedback: many people not seeing value of lecture material • data: room occupancy rates Manipulate Facet Reduce Derive Design & Justification Exercises, • module covers both visualization tooling/code and visualization theory –1 room Change Juxtapose Filter –occupancy measured every 5 min, duration 1 day –lectures: teach theory (assessed with both viz and reasoning) Beyond R • task: characterize space usage pattern • are you coding the right thing? –tutorials: teach tooling/code • derive new data to • how to code it Select Partition Aggregate Tamara Munzner • design show within view –lab 1: 25% mechanics, 49% code, 21% theory, 5% writing Department of Computer Science • propose idioms (visual encoding, interaction) –milestone 1: 5% mechanics, 65% theory, 30% writing • change view over time University of British Columbia • justify idiom choice –milestone 2: 15% mechanics, 45% code, 38% theory, 2% writing • facet across multiple Navigate Superimpose Embed –milestone 3: 5+11=15% mechanics, 10% code, 75% theory views DSCI 532, Data Visualization 2 • today: in-class practice on theory to help you do well on milestone 3 Week 4, Jan 23 / Jan 25 2018 • reduce items/attributes –bar is set considerably higher for milestone 3 than for milestones 1 & 2 www.cs.ubc.ca/~tmm/courses/mds-viz2-17 @tamaramunzner within single view • now that more theory has been covered in class 2 3 4 Consider Cardinality Scenario Consider • what’s the cardinality of the data? • Marshall: 68 cities * 40 years * 4 crime types = 10,880 • data: room occupancy rates • what’s the cardinality of the data? • Wine: 130K * 4 = 650,000 –20 rooms • is a single static chart good enough? • is a single static chart good enough? –measured every 5 min, duration 1 day –spatial (hierarchical), quantitative, categorical, free-form text • should you derive any useful additional data? • should you derive any useful additional data? • task: compare space usage patterns between rooms • what are trade-offs between • design –filtering to see one chart at a time • propose idioms (visual encoding, interaction) –showing all side by side with small multiples • justify idiom choice –superimposing all on top of each other 5 6 7 8 Scenario Consider Scenario Consider • data: room occupancy rates in building • what’s the cardinality of the data? • data: room occupancy rates in building • what’s the cardinality of the data? –1 building: 200 rooms across 4 floors • is a single static chart good enough? –1 building: 200 rooms across 4 floors • is a single static chart good enough? –measured every 5 min, duration 1 day –measured every 5 min, duration 1 year • should you derive any useful additional data? • should you derive any useful additional data? –time series + floor plans –time series + floor plans + room sizes • what are trade-offs between • what are trade-offs between • task: characterize space usage patterns • task: characterize space usage patterns –filtering to see one chart at a time –filtering to see one chart at a time –trends, outliers –trends, outliers –showing side by side with small multiples –showing side by side with small multiples –superimposing on top of each other –superimposing on top of each other • design • design • multi-scale structure to exploit? aggregate, zoom, slice/dice, filter? –propose & justify idioms –propose & justify idioms • multi-scale structure to exploit? aggregate, zoom, slice/dice, filter? • can you normalize the data? should you - always vs on demand? • how to handle multi-scale space and multi-scale time? 9 10 11 12 Idiom: choropleth map Population maps trickiness • use given spatial data • beware! –when central task is understanding spatial • absolute/counts vs normalized/relative relationships • population density vs per capita • data • investigate with Ben Jones Tableau Normalized vs Absolute –geographic geometry Public demo Design Choices –table with 1 quant attribute per region • http://public.tableau.com/profile/ • encoding http://bl.ocks.org/mbostock/4060606 ben.jones#!/vizhome/PopVsFin/PopVsFin 
 (Additional Context) Are Maps of Financial Variables just Population –use given geometry for area mark boundaries Maps? –sequential segmented colormap [more later] • yes, unless you look at per capita –(geographic heat map) (relative) numbers [ https://xkcd.com/1138 ] 13 14 15 16

Idiom: Bayesian surprise maps Idioms: radial bar chart, star plot Axis Orientation • use models of expectations to highlight surprising values • radial bar chart • confounds (population) and variance (sparsity) Rectilinear Parallel Radial –radial axes meet at central ring, line mark • star plot –radial axes, meet at central point, line mark • bar chart Radial vs Rectilinear –rectilinear axes, aligned vertically • accuracy –length unaligned with radial • less accurate than aligned with rectilinear [Surprise! Bayesian Weighting for De-Biasing Thematic Maps. Correll and Heer. Proc InfoVis 2016] [Vismon: Facilitating Risk Assessment and Decision Making In Fisheries Management. Booshehrian, Möller, Peterman, and Munzner. Technical Report TR 2011-04, Simon Fraser University, https://medium.com/@uwdata/surprise-maps-showing-the-unexpected-e92b67398865 https://idl.cs.washington.edu/papers/surprise-maps/ 17 18 19 School of Computing Science, 2011.] 20 "Diagram of the causes of mortality in the army in the Idiom: glyphmaps Radial Orientation: Radar Plots “Radar graphs: Avoid them (99.9% of the time)” East" (1858) • rectilinear good for linear vs nonlinear trends • radial good for cyclic patterns Axis Orientation Rectilinear Parallel Radial [Glyph-maps for Visually Exploring Temporal Patterns in Climate Data and Models. Wickham, LIMITATION: Not good when categories aren’t cyclic http://www.thefunctionalart.com/2012/11/radar-graphs-avoid-them-999-of-time.html Hofmann, Wickham, and Cook. Environmetrics 23:5 (2012), 382–393.] [Slide courtesy of Ben Jones] [Slide courtesy of Ben Jones] [Slide courtesy of Ben Jones] 24 Radial orientation Overview first, zoom and filter, details on demand Thursday Axis Orientation • influential mantra from Shneiderman • perceptual limits • Beyond R Rectilinear –polar coordinate asymmetry –Ana on broader landscape [The Eyes Have It: A Task by Data Type Taxonomy for Information Visualizations. • angles lower precision than lengths Shneiderman. Proc. IEEE Visual Languages, pp. 336–343, 1996.] –Ana on direct comparison of Tableau to R • frequently problematic –Vaden on python interactive tools Query • sometimes can be deliberately exploited! • overview = summary • Evaluations Identify Compare Summarise • for 2 attribs of very unequal importance Evaluations Parallel –microcosm of full vis design problem • Further Design & Justification Exercises • Next Steps Radial [Uncovering Strengths and Weaknesses of Radial Visualizations - an Empirical Approach. Diehl, Beck and Burch. IEEE TVCG 25 26 27 28 (Proc. InfoVis) 16(6):935--942, 2010.] Viz theory How to handle complexity: 4 families of strategies Scenarios last time Scenario • block feedback: many people not seeing value of lecture material • 1 room, occupancy every 5 min over 1 day • data: currency exchange rates Manipulate Facet Reduce Derive • module covers both visualization tooling/code and visualization theory • 20 rooms, occupancy every 5 min over 1 day –30 countries (each against CAD) Change Juxtapose Filter –measured every 5 min, duration 5 years –lectures: teach theory (assessed with both viz and reasoning) • 200 rooms across 4 floors, occupancy every 5 min over 1 day, floor plans –time series + country names + continent names (+ map shapefiles) + country • are you coding the right thing? • 200 rooms, 4 floors, occupancy every 5 min over 1 year, floor plans, 
 populations –tutorials: teach tooling/code room sizes • task: find groups of similarly-performing currencies • derive new data to • how to code it Select Partition Aggregate show within view –lab 1: 25% mechanics, 49% code, 21% theory, 5% writing –milestone 1: 5% mechanics, 65% theory, 30% writing • change view over time –milestone 2: 15% mechanics, 45% code, 38% theory, 2% writing • facet across multiple • design Navigate Superimpose Embed –milestone 3: 5+11=15% mechanics, 10% code, 75% theory views –propose & justify idioms • today: in-class practice on theory to help you do well on milestone 3 • reduce items/attributes –bar is set considerably higher for milestone 3 than for milestones 1 & 2 within single view • now that more theory has been covered in class 29 30 31 32