Strategies For Effective Data Visualization
Strategies for EffectiveData VisualizationAnneli JoplinNovember 8, 2017anneli@rice.edu
Visualization is inherently open-endedSizeShadeAspect idthShapeScaleMotionYX
Best practices depend on context
Approach dictates likelihood of bability of success ctive
An intentional approach to datavisualizationWhat we knowabout perceptionNew ideasto considerData visualizationdesign process
Implications of currentperception research
Vision is a multistep processLight triggers a neural signal through the optic nerveColin Ware, Information Visualization, 2004
Vision is a multistep processBrain identifies basic features first, and then analyzes furtherColin Ware, Information Visualization, 2004
Preattention“When something just catches our eye, it istapping into our earliest stages of attention.”– Stephanie Evergreen, Presenting Data Effectively9
Preattentive processing is instantaneousHow many 254869752158652321221345890011001457Preattentive processing 10 msec per itemTypical processing 40 msec per itemPattern recognitionPreattentive contrastAlberto Cairo, The Functional Art, 2013 / Colin Ware, Information Visualization, 2004
Gestalt principles of pattern recognitionThe visual brain Evolved to detect patterns Groups similar objects Separates different objects
PROXIMITYObjects close together are grouped467281276561271279948278239113576133Alberto Cairo, The Functional Art, 2013459166417183222587771186222177336993
PROXIMITYObjects close together are groupedAlberto Cairo, The Functional Art, 2013
SIMILARITYSimilar objects perceived as a groupAlberto Cairo, The Functional Art, 2013 / USA by Alexander Skowalsky for the Noun Project
Scatter bubble plot encode times in colorBubbles of various sizes are grouped via color (similarity)Forsyth Alexander, When data imitates art, www.ibm.com/blogs/business-analytics
CONNECTEDNESSLinked objects form a natural groupAlberto Cairo, The Functional Art, 2013
CONNECTEDNESSConnecting lines visually identify pairsWhich dots do yougroup and why?
CONNECTEDNESSConnecting lines visually identify pairsConnectednessrelates the two circlesfrom each category
ENCLOSUREEnclosed objects form a natural groupAlberto Cairo, The Functional Art, 2013
ENCLOSUREEnclosed objects are evaluated togetherBang Wong, Nature Methods, Vol. 7 No. 11, 2010
CLOSURETendency to perceive complete formsStephen Few, Show Me the Numbers, 2nd Edition, 2012
CLOSURETendency to perceive complete formsNo need to define area of graph completelyRedundant enclosureintroduces a distractionStephen Few, Show Me the Numbers, 2nd Edition, 2012
SYMMETRYSymmetry suggests a visual wholeColin Ware, Information Visualization, 2004
SYMMETRYButterfly plots highlight differences
CONTINUITYCurved contours imply connectionColin Ware, Information Visualization, 2004
Curved connections are easier to followWhy did we evolve to identify contours?Alberto Cairo, The Functional Art, 2013 / Colin Ware, Information Visualization, 2004ç
Curves help the viewer visually followconnections through crowded dataSocial Networks, behance.net/gallery
Gestalt summaryTake advantage of pattern recognition tendencies mmetryContinuityPattern recognitionPreattentive contrast
Range of evolved preattentive attributesTypeFormColorSpatial closureHueIntensity2D positionStephen Few, Show Me the Numbers, 2nd Edition, 2012
Alter a preattentive attribute to makesomething stand outRolandi, M. et al. Adv. Mater. 2011
Limits to distinct perceptionPreattentive processing limited to 1 attribute at a timeColor intensity onlyStephen Few, Show Me the Numbers, 2nd Edition, 2012Intensity and shape
Overwhelming repetition results in lossof meaningToo many bright colorsmeans nothing stands outMartin Krzywinski, Nature Methods, Vol. 10 No. 5 2013
Natural scenes exhibit muted colorsReserve bright colors for emphasisBrightmetrics, Using Color in Data Visualization, 2010
Above all else, show the dataTUFTEData ink ratio data inktotal ink used in the graphicThe Visual Display ofQuantitative InformationEdward Tufte, The Visual Display of Quantitative Information, 1983
Clutter distracts from preattentive cuesRemove all chartjunk, for example: Distracting patterns Gridlines Elements only for “artistic appeal”Edward R. Tufte, The Visual Display of Quantitative Information, 1983
3D effects are almost always chartjunkNils Gehlenborg and Bang Wong, Nature Methods, Vol. 9 No. 9 2012
Visually separate data from otherelementsSimilarity between theellipses and linesreduces visual contrastMartin Krzywinski, Nature Methods, Vol. 10 No. 3 2013
“Sometimes clarity demands morespace” – Stephen FewBEFORESeparating traces intotrellis display highlightstrends more effectivelyAFTER
Make emphasis more effective byeliminating excess decorationWhat would you removefrom this chart?Size Matters, rs/
Preattentive contrast summaryLimit preattentive attributes to emphasis Rely on muted colors Soften gridlines, axes, labels, etc. Remove chartjunk
Visual information requires decodingColin Ware, Information Visualization, 2004
Visual information requires decodingImplications for data visualization –1. Working memory limits number of items remembered2. Perception accuracy is distance dependent3. Accuracy of perception influenced by visuals
Keep the number of items displayed inone visualization to 4 if possible
Reduce distance between comparabledata to increase accuracyMarc Streit and Nils Gehlenborg , Nature Methods, Vol. 11 No. 2 2014
Select attribute based on purposeData typesCategoricalQuantitativeDivide informationMeasure thingsCompanyParticipantMoleculeOrder (1, 2, 3)AddressTimeCountIntensityProfit
Few attributes can encode quant. dataTypeFormColorSpatial closureHueIntensity2D positionStephen Few, Show Me the Numbers, 2nd Edition, 2012Quantitative?YesYes, but limitedNoYes, but limitedNoNoNoYes, but limitedYes
Shifts in color are not visually equivalentto changes in valueCommonly utilized color scales are not perceived accuratelyBang Wong, Nature Methods, Vol. 8 No. 3 20111
Perception of color depends onsurroundingsUse color for labeling, emphasis or when value doesn’t matterBang Wong, Nature Methods, Vol. 7 No. 8 2010
Length is perceived quantitativelyNumber and visual length are tied togetherThis works to ouradvantage in a barchart, for example
Bar charts must start at zeroLength has an inherent numerical valueScale 0 – 100Scale 60 – 100Data encoded with length is highlydistorted with a shortened scale
An alternative – the dot plot2D position does not elicit a numerical valueScale 0 – 1002D positiondoes notrequire a 0value forquantitativecomparisonScale 60 – 100
Dot plots display multiple data setsmore clearly than bar charts
Lollipop charts also compare valueswithout emphasizing lengthBar chart with lessemphasis on length
Cleveland and McGill identified 10elementary perceptual tasksWilliam Cleveland, Graphical Perception, 1984
Graphical perception attributes inorder of accuracyAllows moreaccuratejudgmentsposition along a common scaleposition along nonaligned scaleslengthangleareaAllows moregenericjudgmentsvolume*Accuracy is notalways better,just make intentionalchoices based onpurposecurvatureshading, color saturationWilliam Cleveland, The Elements of Graphing Data, 1994 / Alberto Cairo, The Functional Art
Bar charts are easier to evaluateaccurately than pie chartsPosition along common scale area or angle
Simple bar charts more accurate thanstacked barsPosition along common scale length
Use small multiples instead of stackedbars when numbers matterRetains common axis, butalso enables comparison
Curve comparisons are difficult, plotdifference insteadCurves A and BDifference (A – B)
Select an aspect ratio that places keylines close to 45 Angles around 45 areperceived accuratelySmall angles are moredifficult to assessNaomi B. Robbins, Creating More Effective Graphs, 2005
Aspect ratio affects perception of dataHow to select the aspect ratio that allows for accuratejudgment?Naomi B. Robbins, Creating More Effective Graphs, 2005
Rescale line graph segments in multiplepanels to improve angle perceptionGregor McInerny, Martin Krzywinski, Nature Methods, Vol. 12 No. 7 2015
Graphical perception summarySelect encoding attributes based on purpose Perception accuracy is distance dependent Position on a common axis perceived most accurately Bar graphs outperform pie charts Small multiples outperform stacked bars Curve perception is not accurate Angles close to 45 are evaluated most easily
Exercise 1 –Accounting for perception
How would you apply visual perceptionprinciples to improve this chart?Example curated by Melissa Clarkson, melissaclarkson.com
One solution – Bar chart trellis displayallows comparison across samplesRedesign created by Melissa Clarkson, melissaclarkson.com
How would you apply visual perceptionprinciples to improve this chart?Example curated by Melissa Clarkson, melissaclarkson.com
One solution – Dot plot allows easycomparison across conditionsRedesign created by Melissa Clarkson, melissaclarkson.com
Strategies to facilitateeffective data visualization
Field of data visualizationTUFTEThe Visual Display ofQuantitative InformationHOLMESDesigner’s Guide to CreatingCharts and Diagrams
Tufte prioritized function, Holmes formTUFTEHOLMESVSNigel Holmes, Designer’s Guide to Creating Charts and Graphs, 1984
Approach style guidelinesProbability of success intentionalapproachEffectiveIneffective
Recommended design process1. Explore data visually2. Identify visualization message3. Select a chart type and create4. Evaluate and revise5. Take advantage of templates
Recommended design process1. Explore data visually2. Identify visualization message3. Select a chart type and create4. Evaluate and revise5. Take advantage of templates
Scatter plot matrixStephen Turner, Scatterplot Matrices in R, 2011, r-bloggers.com
Streamline with a visualization dashboardPreset charts provide an instant view of new dataAlberto Cairo, thefunctionalart.com, 2017
Add interactive components to quicklyfilter and display dataAlberto Cairo, thefunctionalart.com, 2017
Exploring high dimensional dataOnline data display1. Embedding projectorProjectionRice Visualization Lab(closed for relocation)2. HypertoolsTensorFlow, Embeddings, tensorflow.org, 2017 / Andrew C. Heuser, Hypertools, 2017
Recommended design process1. Explore data visually2. Identify visualization message3. Select a chart type and create4. Evaluate and revise5. Take advantage of templates
Evaluate visualizations on bothinformative and emotive lnessIntuitivenessAestheticsEngagementVery usefulUselessAll relevant dataNo relevant dataClear and easyUnclear and difficultInaccurateAccurateUnfamiliarFamiliar, easy to readUglyDistracts from dataPleasing to the eyeNeutralStephen Few, Perceptual Edge, Visual Business Intelligence Newsletter, 2017BeautifulDraws one in
Exercise 2 –Evaluating visualizations
Example 1 – ScatterplotJonathan A. Schwabish, An Economist’s Guide to Visualizing Data, 2014
Coloring and labeling key datafacilitates interpretationJonathan A. Schwabish, An Economist’s Guide to Visualizing Data, 2014
Example 2 – Bar / dot plotJonathan A. Schwabish, An Economist’s Guide to Visualizing Data, 2014
Horizontal dot plot visually comparescategories at two points in timeJonathan A. Schwabish, An Economist’s Guide to Visualizing Data, 2014
Example 3 – Stacked bar chartCastro-Nallar, E. et al Peer J, 2015, accessed at peerj.com/articles/1140/
Recommended design process1. Explore data visually2. Identify visualization message3. Select a chart type and create4. Evaluate and revise5. Take advantage of design templates
Create templates to save timeTemplates eliminate mundane design decisionsSpreadsheets(Excel, Origin)Secondary components(Illustrator, InDesign, PowerPoint)Dashboards(Excel, Tableau)Scripts(Matlab, Python, Origin)
Resources on campusDigital Media CommonsRice Visualization LabCWOVC online resourcesGIS Data CenterCenter for ResearchComputing
Exploring the frontiersof data visualization
Less common chart types provide newmeans of data explorationSeverino Ribecca, The Data Visualization Catalogue, dataviscatalogue.com/methods/
Raw graphs – a free way to experimentwith less common visualizationsSupports conventional andunconventional chart typesrawgraphs.io
Sunburst diagramCapable of displaying hierarchies over multiple levelsUse to show subdivisionsof a multi-level structureSeverino Ribecca, The Data Visualization Catalogue, dataviscatalogue.com/methods/
Sunburst diagram applied to visualizememory usage
Butterfly plot
Deviation bar chart
Slopegraph utilizes angle to comparechange between groupsDisplays changeacross categoriesusing slopeAxes labels alsoserve as ranked listsChart from the New York Times / h/infant stats.html
Parallel coordinates showcase trendsacross dimensionsExtension of slopegraph to high dimensional dataOrder matters – place thedimensions you aim tocompare close togetherProtovis, A Graphical Toolkit for Visualization, http://mbostock.github.io/protovis/ex/cars.html
Brushing highlights relevant data rangesInteractivity allows exploration of trends in the dataSelect a category rangeRobert Kosara, Parallel Coordinates, eagereyes.org
Heat mapUse to highlightoverall data trendsSeverino Ribecca, The Data Visualization Catalogue, dataviscatalogue.com/methods/
Expanded heat map presents largescale patterns in a compact wayStatistical Computing and Graphics Newsletter, Volume 20, December 2009
Radar chartsRepresent the value of multiple variables as a polygonSeverino Ribecca, The Data Visualization Catalogue, dataviscatalogue.com/methods/
Sunburst style chart shows cyclicrelationship*good for high dimensional data*Moritz Stefaner, The Rhythm of Food, truth-and-beauty.net
Horizon chart – high dimensional areachartValue encoded incolor and intensityStephen Few, Time on the Horizon, Perceptual Edge, 2008
Horizon chart – high dimensional areachartTiers grouped toimprove perceptionof differencesStephen Few, Time on the Horizon, Perceptual Edge, 2008
Horizon chart – high dimensional areachartStephen Few, Time on the Horizon, Perceptual Edge, 2008
Horizon chart – high dimensional areachartCollapsed stackspresent compactinformationStephen Few, Time on the Horizon, Perceptual Edge, 2008
Horizon chart – high dimensional areachartHigh intensitypockets stand outStephen Few, Time on the Horizon, Perceptual Edge, 2008
Learn more at cwovc.rice.edu
Stephen Few, Perceptual Edge, Visual Business Intelligence Newsletter, 2017 . Exercise 2 – Evaluating visualizations . Example 1 – Scatterplot Jonathan A. Schwabish, An Economist’s Guide to Visualizing Data, 2014 . Coloring and labeling key data facilitates interpretation
Bruksanvisning för bilstereo . Bruksanvisning for bilstereo . Instrukcja obsługi samochodowego odtwarzacza stereo . Operating Instructions for Car Stereo . 610-104 . SV . Bruksanvisning i original
10 tips och tricks för att lyckas med ert sap-projekt 20 SAPSANYTT 2/2015 De flesta projektledare känner säkert till Cobb’s paradox. Martin Cobb verkade som CIO för sekretariatet för Treasury Board of Canada 1995 då han ställde frågan
service i Norge och Finland drivs inom ramen för ett enskilt företag (NRK. 1 och Yleisradio), fin ns det i Sverige tre: Ett för tv (Sveriges Television , SVT ), ett för radio (Sveriges Radio , SR ) och ett för utbildnings program (Sveriges Utbildningsradio, UR, vilket till följd av sin begränsade storlek inte återfinns bland de 25 största
Hotell För hotell anges de tre klasserna A/B, C och D. Det betyder att den "normala" standarden C är acceptabel men att motiven för en högre standard är starka. Ljudklass C motsvarar de tidigare normkraven för hotell, ljudklass A/B motsvarar kraven för moderna hotell med hög standard och ljudklass D kan användas vid
LÄS NOGGRANT FÖLJANDE VILLKOR FÖR APPLE DEVELOPER PROGRAM LICENCE . Apple Developer Program License Agreement Syfte Du vill använda Apple-mjukvara (enligt definitionen nedan) för att utveckla en eller flera Applikationer (enligt definitionen nedan) för Apple-märkta produkter. . Applikationer som utvecklas för iOS-produkter, Apple .
2.1 Data Visualization Data visualization in the digital age has skyrocketed, but making sense of data has a long history and has frequently been discussed by scientists and statisticians. 2.1.1 History of Data Visualization In Michael Friendly's paper from 2009 [14], he gives a thorough description of the history of data visualization.
language express all the facts in the set of data, and only the facts in the data. Effectiveness A visualization is more effective than another visualization if the information conveyed by one visualization is more readily perceived than the information in the other visualization. Design Principles [Mackinlay 86]
discussing the challenges of big data visualization, and analyzing technology progress in big data visualization. In this study, authors first searched for papers that are related to data visualization and were published in recent years through the university library system. At this stage, authors mainly summarized traditional data visualization
