Visual Analytics Review IAT 355 Lyn Bartram Overview
Visual Analytics Review IAT 355 Lyn Bartram
Overview • Topics ( in no particular order) ■ ■ ■ ■ Data models and analytics Information visualization techniques: Types and components Interaction Perception Cognition Navigation and Scent Presentation and screen space IAT 355 Introduction
Overview and definitions IAT 355 Introduction
Information visualization • visual metaphors for non-inherently spatial data such as the exploration of text-based document databases. ■ More abstract • Assign structure and position to information that has none • • • Text Statistics Finance/Business Internet Software IAT 355 Introduction
Visual analytics • analytical reasoning supported by the interactive visual interface • Intersection of visualization with data analysis • Biology • National security IAT 355 Introduction
Visual thinking involves: • Constructing visual queries on displays • Visual search strategies through eye movements and attention to relevant patterns • Visual notification and attention “redirection” to new patterns and events • Well structured balance of elements and tasks IAT 355 Introduction
Data Analytics IAT 355 Introduction
Data We Use • Data Models • Descriptive Statistics • Types • Distribution • Metadata • Clusters • Aggregates Show Me the Numbers! : Data
Data models • take raw data and transform it into a form that is more workable • Main idea: build a model ■ Individual items are called cases or records ■ Cases have attributes : an attribute is a value of a variable or factor ■ In vis terms, a dimension
How many dimensions? • Data sets of dimensions 1, 2, 3 are common • • • Number of variables per class 1 - Univariate data 2 - Bivariate data 3 - Trivariate data >3 - Hypervariate data ■ These are the fun and interesting ones! But hard! Show Me the Numbers! : Data
Data Types (measurements) • Nominal: categorical, ( equal or not equal to other values) ■ Example: gender, Student Number ■ No concept of relative relation other than inclusion in the set • Ordinal : sequential ( obeys < > relation, ordered set ■ Example: Size of car, speed settings on road ■ Example: mild, medium, hot, suicide ■ Distance is not uniform Show Me the Numbers! : Data
Data Types 2 • Interval : Relative measurements, no fixed zero point. ■ Data is numerical, not categorical. Rank order among variables is explicit with an equal distance between points in the data set: -2, -1, 0, +1, +2 ■ can say “twice as much as” ■ Example: height above sea level, hours in a day • Ratio: Interval data with absolute zero ■ Example: account balance, degrees Kelvin Show Me the Numbers! : Data
Dimensions • Data Dimensions are classified as: ■ Quantitative i. e. numerical • Continuous (e. g. p. H of a sample, patient cholesterol levels) • Discrete (e. g. number of bacteria colonies in a culture) ■ Categorical • Nominal (e. g. gender, blood group) • Ordinal (ranked e. g. mild, moderate or severe illness). Often ordinal variables are re-coded to be quantitative. 13
Metadata Mary Tom Louise 65432101 98765651 89846251 20 22 19 Sep 2006 Jan 2004 Sep 2005 4. 0 2. 3 3. 04 • Descriptive information about the data • Might be something as simple as the type of a variable, or could be more complex ■ For times when the table itself just isn’t enough ■ Example: if variable 1 is “l”, then variable 3 can only be 3, 7 or 16 • Missing values, uncertainty or importance are all examples of metadata Show Me the Numbers! : Data
Primary types of data analysis • Qualitative • Descriptive. Used to describe the distribution of a single variable or the relationship between two nominal variables (mean, frequencies, cross-tabulation) • Inferential (Used to establish relationships among variables; assumes random sampling and a normal distribution) • Nonparametric (Used to establish causation for small samples or data sets that are not normally distributed) Show Me the Numbers! : Data
Descriptive Statistics • Range • Min/Max • Average • Median • Mode Show Me the Numbers! : Data Distribution Statistics • Variance • Error • Standard Deviation • Histograms and Normal Distributions
Range, Min, Max • The Range ■ Difference between minimum and maximum values in a data set ■ Larger range usually (but not always) indicates a large spread or deviation in the values of the data set. (73, 66, 69, 67, 49, 60, 81, 78, 62, 53, 87, 74, 65, 74, 50, 85, 45, 63, 100)
Average = measure of centrality • Measures of location indicate where on the number line the data are to be found. Common measures of location are: • (i) the Arithmetic Mean, • (ii) the Median, and • (iii) the Mode
The mean is vulnerable to problems 0 2. 5 7. 5 10 4. 8 0 2. 5 4. 8 The data may or may not be symmetrical around its average value
• The Median ■ The middle value in a sorted data set. Half the values are greater and half are less than the median. ■ Another measure of central location in the data set. (45, 49, 50, 53, 60, 62, 63, 65, 66, 67, 69, 71, 73, 74, 78, 81, 85, 87, 100) Median: 68 (1, 2, 4, 7, 8, 9, 9)
• The Median ■ May or may not be close to the mean. ■ Combination of mean and median are used to define the skewness of a distribution. 0 Show Me the Numbers! : Data 2. 5 7. 5 6. 25 10
The Mode • The Mode ■ The most frequent occurring value. ■ Another measure of central location in the data set. ■ (45, 49, 50, 53, 60, 62, 63, 65, 66, 67, 69, 71, 73, 74, 78, 81, 85, 87, 100) ■ Mode: 74 ■ Generally not all that meaningful unless a larger percentage of the values are the same number Show Me the Numbers! : Data
When do we use what? • Dependent on how the data are distributed ■ Note if mean=median=mode then the data are said to be symmetrical • Rule of thumb: ■ use mean if data are normally distributed and variance is within constraints ■ Use median to reduce effects of outliers Show Me the Numbers! : Data
Summary http: //statistics. laerd. com/statistical-guides/measures-central-tendency-mean-modemedian. php Show Me the Numbers! : Data
Data distribution • • Measures of dispersion characterise how spread out the distribution is, i. e. , how variable the data are. Commonly used measures of dispersion include: 1. 2. 3. 4. Show Me the Numbers! : Data Range Variance & Standard deviation Coefficient of Variation (or relative standard deviation) Inter-quartile range
Measures of variance • Variance ■ One measure of dispersion (deviation from the mean) of a data set. The larger the variance, the greater is the average deviation of each datum from the average value • Standard Deviation ■ the average deviation from the mean of a data set. • An outlier is an datum which does not appear to belong with the other data Show Me the Numbers! : Data
Inter-quartile range • The Median divides a distribution into two halves. • The first and third quartiles (denoted Q 1 and Q 3) are defined as follows: ■ 25% of the data lie below Q 1 (and 75% is above Q 1), ■ 25% of the data lie above Q 3 (and 75% is below Q 3) • The inter-quartile range (IQR) is the difference between the first and third quartiles, i. e. IQR = Q 3 - Q 1 27
Box-plots • A box-plot is a visual description of the distribution based on ■ ■ ■ Minimum Q 1 Median Q 3 Maximum If a data point is < lower limit or > upper limit, the data point is considered to be an outlier. • Useful for comparing large sets of data 28
Distribution is important for Aggregation • Visualization helps us see relations – or the trends of them - as visual patterns • a lot of what we visualize are the descriptive statistics ■ Example: mean income vs median income ■ Need to ensure that the univariate units of visualization are legit • Rule: check your core units /variables. If hey are descriptive, look at the distribution Show Me the Numbers! : Data
Example: job losses in US over time Show Me the Numbers! : Data
Example: job losses in US over time Show Me the Numbers! : Data
Show Me the Numbers! : Data
2 D Visualization Classes IAT 355 Introduction
Types of Symbolic Displays (Kosslyn 89) • Graphs • Maps • Charts • Diagrams
Types of Symbolic Displays • Graphs ■ at least two scales required ■ values associated by a symmetric “paired with” relation • Examples: scatter-plot, bar-chart, layer-graph
Graphs • • Encode quantitative information using position and magnitude of geometric objects. Examples: scatter plots, bar charts.
Types of Symbolic Displays • Charts ■ discrete relations among discrete entities ■ structure relates entities to one another ■ lines and relative position serve as links • Examples: ■ Family tree ■ Flow chart ■ Network diagram
Map • Internal relations determined (in part) by the spatial relations of what is pictured ■ Grid: geometric metadata • Locations identified by labels • Nominal metadata • Examples: • Map of census data • Topographic maps Jan 21, 2011 IAT 355 38
Choropleth Map • Areas are filled and colored differently to indicate some attribute of that region Jan 21, 2011 IAT 355 39
Diagrams • Schematic pictures of objects or entities • Parts are symbolic (unlike photographs) ■ how-to illustrations ■ figures in a manual From Glietman, Henry. Psychology. W. W. Norton and Company, Inc. New York, 1995
Graph Components • Framework (spatial substrate) ■ Measurement types, scale ■ Geometric Metadata • Content ■ Marks, lines, points ■ Data • Labels ■ Title, axes, ticks ■ Nominal Metadata Jan 21, 2011 IAT 355 41
Marks • Things that occur in space ■ ■ Points Lines Areas Volumes Jan 21, 2011 IAT 355 42
Graphical Properties • Size, shape, color, orientation. . . Jan 21, 2011 IAT 355 Spatial Properties Object Properties Expressing Extent Position, Size Greyscale Differentiating Marks Orientation Color, Shape, Texture 43
What goes where • In univariate representations, we often think of the data case as being shown along one dimension, and the value (quantity) in another Y Axis is quantitative Graph shows change in Y over continuous range X Graph shows value of Y for 4 cases Jan 21, 2011 IAT 355 44
Bivariate Data • Representations ■ Scatter plot Price ■ Each mark is a data case ■ Want to see relationship between two variables ■ What is the pattern? ■ Note both variables are continuous data Jan 21, 2011 IAT 355 Mileage 45
Multivariate: Project data onto other graphical variables • E. G. , Use blob attribute for another variable Price Mileage Jan 21, 2011 IAT 355 46
Alternative • Represent each variable on its own line Small multiples Jan 21, 2011 IAT 355 47
Data projection • Fundamentally, we have 2 display dimensions • For data sets with >2 variables, we must project data down to 2 D ■ Come up with visual mapping that locates each dimension into 2 D plane • Computer graphics 3 D->2 D projections IAT 355: Mutivariate Data
What is Multivariate Data? • Each data point has N variables or observations • Each observation can be: ■ nominal or ordinal ■ discrete or continuous ■ scalar, vector, or tensor • May or may not have spatial, temporal, or other connectivity attribute This slide courtesy of Matt Ward, UC Berkeley
Methods for Visualizing Multivariate Data • Dimensional Subsetting • Dimensional Reorganization dimensional re-ordering • Dimensional Embedding • Dimensional Reduction This slide courtesy of Matt Ward, UC Berkeley
Dimensional Subsetting • Scatterplot matrix displays all pairwise plots • Selection allows linkage between views • Clusters, trends, and correlations readily discerned between pairs of dimensions • Small mulitples UC Berkeley, 09/19/00
Dimensional Reorganization • Parallel Coordinates creates parallel, rather than orthogonal, dimensions. • Data point corresponds to polyline across axes • Clusters, trends, and anomalies discernable as groupings or outliers, based on intercepts and slopes UC Berkeley, 09/19/00
Dimensional Reorganization (2) • Glyphs map data dimensions to graphical attributes • Size, color, shape, and orientation are commonly used • Similarities/differences in features give insights into relations UC Berkeley, 09/19/00
Dimensional Embedding • Dimensional stacking divides data space into bins • Each N-D bin has a unique 2 -D screen bin • Screen space recursively divided based on bin count for each dimension • Clusters and trends manifested as repeated patterns UC Berkeley, 09/19/00
Dimensional Reduction • Map N-D locations to M-D display space while best preserving N-D relations • CLUSTERING • Approaches include MDS, PCA, and Kohonen Self Organizing Maps • Relationships conveyed by position, links, color, shape, size, etc. UC Berkeley, 09/19/00
Perception IAT 355 Introduction
Preattentive processing • A limited set of visual properties are processed preattentively (without need for focusing attention). ■ Visual features • This is important for the design of visualizations ■ ■ what can be perceived immediately what properties are good discriminators what can mislead viewers Differentiate items “at a glance” – THEY POP OUT Some examples from Chris Healey: http: //www. csc. ncsu. edu/faculty/healey/PP/PP. html IAT 355 Perception
Preattentive processing features • Form ■ ■ ■ ■ Line orientation Line length line width Size Curvature Spatial grouping Blur numerosity IAT 355 Perception • Colour ■ Hue ■ Intensity • Motion ■ Flicker ■ Direction of motion • Spatial position ■ 2 D position ■ Stereo depth ■ Concavity/convexity shape from shading
Coding with several features: conjunction • What happens with more complex patterns ? ■ a large red circle, not just something that is red or something that is large? • slow if the surrounding objects are large (but not red ones) and other red sizes. ■ a serial search of either the red or the large circles. • conjunction search - searching for the specific conjunction of colour and size attributes. ■ generally not pre-attentive, although there a few very interesting exceptions. IAT 355 Perception
Conjunction does not pop out IAT 355 Perception
Compound features do not pop out IAT 355 Perception
Surrounded colours do not pop out IAT 355 Perception
Similarity and integral dimensions separable • • integral A: separable dimensions allow both groupings to be perceived – but not simultaneously B: with integral dimensions we can see both to construct a grid Patterns | IAT 355 | 25. 03. 2012
Similarity and the separability of dimensions Integral dimensions (colour and grayscale) are used to delineate rows and columns Separable dimensions (colour and texture) make it easier to attend separately to either the rows or the columns
Integral-Separable • Not one or other, but along an axis
Integral vs. Separable Dimensions Integral Separable Jan 17, 2011 [Ware 2000] 66
Colour
Perceptual issues • Contrast effects • Luminance and brightness • Colour deficiency IAT 355 Introduction
Simultaneous contrast effects • a gray patch placed on a dark background looks lighter than the same gray patch on a light background. • http: //www. michaelbach. de/ot/lum_dynsimcontrast/index. html Contrast, Luminance and Colour | IAT 814 | 21. 09. 2011
Effects cause error! • Simultaneous contrast effects can result in large errors of judgment when reading quantitative (value) information displayed using a gray scale. • Ware et al showed an average error of 20% of the entire gray scale in a map encoding gravity fields using 16 levels of gray. Contrast, Luminance and Colour | IAT 814 | 21. 09. 2011
Crispening Contrast, Luminance and Colour | IAT 814 | 21. 09. 2011
What about colour? • Colour perception is relative • We are sensitive to small differences ■ hence need sixteen million colours • Not sensitive to absolute values ■ hence we can only use < 10 colours for coding Contrast, Luminance and Colour | IAT 814 | 21. 09. 2011
Vischeck • Simulates color vision deficiencies ■ Web service or Photoshop plug-in ■ Robert Dougherty and Alex Wade • www. vischeck. com Deuteranope Contrast, Luminance and Colour | IAT 814 | 21. 09. 2011 Protanope Tritanope
Fundamental Uses • • • To label (colour as noun) To measure ( colour as quantity/value) To represent (colour as representation) ■ to imitate reality • To enliven or decorate (colour as beauty) IAT 355 | Colour for Information Display
Colour great for classification • • • Rapid visual segmentation Colour helps us determine type Only about six categories Information Visualization Colin Ware IAT 355 | Colour for Information Display
Contrast Creates Pop-out Hue and lightness IAT 355 | Colour for Information Display Lightness only
Pop-out vs. Distinguishable • Pop-out ■ Typically, 5 -6 distinct values simultaneously ■ Up to 9 under controlled conditions • Distinguishable ■ 20 easily for reasonable sized stimuli ■ More if in a controlled context ■ Usually need a legend IAT 355 | Colour for Information Display
Data to Color • Types of data values ■ Nominal, ordinal, numeric ■ Qualitative, sequential, diverging • Types of color scales ■ Hue scale • Nominal (labels) • Cyclic (learned order) ■ Lightness or saturation scales • Ordered scales • Lightness best for high frequency • More = darker (or more saturated) • Most accurate if quantized IAT 355 | Colour for Information Display Quantized • Signal varies continuously Discretized • Restricted to a prescribed set of values
Pseudocoloring • • Pseudocoloring is the technique of representing continuously varying map values with a sequence of colours Sometimes overlaid on luminosity information ■ Need to use an isoluminant color map to avoid distortion • • “intuitive” based on lightness, saturation No perceptually based hue scales ■ Need to be learned IAT 355 | Colour for Information Display
Pseudocoloring IAT 355 | Colour for Information Display
Thematic Maps US Census Map IAT 355 | Colour for Information Display Mapping Census 2000: The Geography of U. S. Diversity
How many dimensions? • Univariate scale is a path in a colour space ■ Progression along a line • Multivariate is: ■ Plane? 2 D ■ Volume ? 3 D ■ Rules for color mixing • Only perceptual coding is 2 D ■ lightness x saturation • Color for multivariate only works well for highly quantized data ■ Like a mnemonic for a labeling scheme • IAT 355 | Colour for Information Display
Brewer System IAT 355 | Colour for Information Display http: //www. colorbrewer. org
What Defines Layering? • Perceptual features ■ Contrast (especially lightness) ■ Color, shape and texture • Task and attention ■ Attention affects perception • Display characteristics ■ Brightness, contrast, “gamma” IAT 355 | Colour for Information Display Emergency
General guidelines … or from Tufte to practice [Stone, Ware] • Assign colour according to function • Use contrast to highlight • Use analogy to group • Control value contrast for legibility • Break isoluminance with borders IAT 355 | Colour for Information Display
Visual Organisation and Patterns IAT 355 Introduction
Pattern learning • People who work with visualizations must learn the skill of seeing patterns in data. • In terms of making visualizations that contain easily identified patterns, one strategy is to rely on pattern-finding skills that are common to everyone. • Good idea to use priming to enhance perceptual receptivity Patterns | IAT 355 | 25. 03. 2012
The Gestalt laws The core laws 1. Proximity 2. Similarity 3. connectedness 4. Continuity 5. Symmetry 6. Closure 7. Relative Size 8. Common fate Patterns | IAT 355 | 25. 03. 2012 Principal effects 9. Figure - ground 10. Prägnanz : the “organising principle”
Proximity: design implications • Emphasize relationship by proximity Patterns | IAT 355 | 25. 03. 2012 • Emphasize relationship by spatial density
Similarity • Similarity between the elements in alternate rows causes the row percept to dominate Patterns | IAT 355 | 25. 03. 2012
Continuity • • • The Gestalt principle of continuity states that we are more likely to construct visual entities out of objects that are smooth and continuous, rather than those that contain abrupt changes in direction. We see a-b crossing c-d not a-d or b-c Patterns | IAT 355 | 25. 03. 2012 A D C B
Connectedness • Connecting graphical objects by a line is a very powerful way of expressing that there is a relationship between them • Basis of node-link diagrams • Most common method of indicating relationships Patterns | IAT 355 | 25. 03. 2012
Symmetry • Symmetry creates visual whole • Powerful organising principle • b and c are seen as figures/objects, where a is a pair of parallel lines • We construct objects in the world Patterns | IAT 355 | 25. 03. 2012 (a) (b) (c)
Closure • • • Over-rules proximity ! A closed contour tends to be seen as an object The Gestalt psychologists argued that there is a perceptual tendency to close contours that have gaps a circle behind a rectangle as in (a), not a broken ring as in (b). Patterns | IAT 355 | 25. 03. 2012
Figure and Ground • Confronted by a visual image, we seem to need to separate a dominant shape (a 'figure' with a definite contour) from what our current concerns relegate to 'background' (or 'ground') • Symmetry, white space, and closed contour contribute to perception of figure. • The perception of figure as opposed to ground can be thought of as the fundamental perceptual act of identifying objects. Patterns | IAT 355 | 25. 03. 2012
Prägnanz • A stimulus will be organized into as good a figure as possible. Here, good means symmetrical, simple, and regular. • here we see a square overlapping a triangle, not a combination of several complicated shapes. Patterns | IAT 355 | 25. 03. 2012
IAT 355 Introduction
Cognitive tasks
Low-Level Components of Analytic Activity in Information Visualization Amar, Eagan, and Stasko • Identified 10 low-level analysis tasks that largely capture people’s activities while employing information visualization tools for understanding data • • • Retrieve value Filter Compute derived value Find extrema Sort • • • Determine range Characterize distribution Find anomalies Cluster Correlate 99
Example Visual Analytics Characteristics • Whole-part relationship: multiple levels of information extraction • Relationship discovery: high dimensional analytics to detect the expected and discover the unexpected • Combined exploratory and confirmatory analytics • Selection, search (bool. and similarity) and groupings • Temporal and geospatial analytics • Extensive labeling: everything active on screen • Multiple linked views • Analytic interactions are foundational to critical thinking • Analytic reasoning framework • Capture analytic snippets for reporting • Both general and application specific applications 100
Understanding • People utilize an internal model that is generated based on what is observed • Tversky calls the internal model a cognitive map • Just don’t have one big one • Have large number of these for all different kinds of things • Collection of cognitive maps --> Cognitive collage
Process model 1: Navigation (spence) Content Model Browse Cognitive map Browsing strategy Formulate a browsing strategy Internal model New view Interpretation Interpret
Process model 2: Knowledge Crystallization Overview Zoom Filter Details-on-demand Browse Search query Task Author, decide or act Forage for data Search for schema Reorder Cluster Class Average Promote Detect pattern Abstract Extract Compose Problem-solve Instantiate schema Instantiate Read fact Read comparison Read pattern Manipulate Create Delete
Process task Raw data Data tables Data transformations Visual Structures Visual mappings View transformations
Visual Task Taxonomy (Zhou & Feiner) # Relational tasks Associate -- Collocate -- Connect -- Unite -- Attach Background Categorize -- Mark. Distribute Cluster -- Outline -- Individualize Compare -- Differentiate -- Intersect Correlate -- Plot -- Mark. Compose Distinguish -- Mark. Distribue -- Isolate Emphasize -- Focus -- Isolate -- Reinforce Generalize -- Merge Identify -- Name -- Portray -- Individualize -- Profile Locate # Direct visual -- Position # organizing and -- Situate # encoding tasks -- Pinpoint Encode -- Outline -- Label Rank -- Symbolize -- Time -- -- Quantify Reveal -- -- Iconify -- Expose -- Portray -- Itemize -- Tabulate The nested items are refinements -- Specify -- Plot of particular ways of achieving -- Structure task -- Separate Switch -- Trace -- Map
Dimensions • Visual tasks have two main dimensions 1. 2. Visual accomplishments - describe presentation intents that task might help to achieve Visual implications - particular type of visual action that visual task may carry out
1. Visual Accomplishments • All about presentation intent • Classified into two categories: ■ Tasks that inform the user (e. g. , make a presentation with ppt) ■ Tasks that enable user to explore or compute (e. g. , decide which stock to buy) • Each of these can be broken down further
Visual Accomplishments Inform Elaborate Summarize Emphasize Reveal Associate Background Categorize Cluster Compare Correlate Distinguish Generalize Identify Locate Rank Enable Compute Explore Search Categorize Cluster Compare Correlate Distinguish Emphasize Identify Locate Rank Reveal Verify Sum Categorize Compare Correlate Distinguish Identify Locate Rank Reveal Differentiate Correlate Locate Rank
2. Visual Implications • Categorize various visual tasks by whether they imply ■ Certain types of visual organization ■ Certain ways of visual signaling ■ Certain paths of visual transformation
interaction
Shneiderman’s Taxonomy of Information Visualization Data Types • • 1 -D Linear 2 -D Map 3 -D World Multi-Dim • Temporal • Tree • Network interaction | IAT 355 Document Lens, See. Soft GIS, Medical imagery CAD, Medical, Molecules, Architecture Parallel Coordinates, Spotfire, Influence Explorer, Table. Lens Perspective Wall, Life. Lines, Lifestreams Cone/Cam/Hyperbolic, Tree. Browser, Treemap Netmap, net. Viz, Multi-trees
Shneiderman’s Taxonomy of Information Visualization Tasks • • Overview: see overall patterns, trends Zoom: see a smaller subset of the data Filter: see a subset based on values, etc. Detailed on demand: see values of objects when interactively selected • Relate: see relationships, compare values • History: keep track of actions and insights • Extract: mark and capture data interaction | IAT 355
Interaction Techniques • • View Specification (map data to visual variables) Navigation (pan, zoom, scale, rotate) Selection Highlighting (Brushing) Filtering Sorting Extract Data interaction | IAT 355
Advanced Interaction Techniques • Brushing and Linking • Overview + Detail • Focus + Context • Panning and Zooming interaction | IAT 355
How long in majors Brushing and Linking • • Select (“brush”) a subset of data See selected data in other views • The components must be linked by tuple (matching data points), or by query (matching range or values) interaction | IAT 355 salaries
Filtering: Dynamic Queries • Spotfire, by Ahlberg & Shneiderman ■ http: //hcil. cs. umd. edu/video/1994_visualinfo. mpg ■ Now a very sophisticated product: • http: //spotfire. tibco. com/products/gallery. cfm interaction | IAT 355
Interaction strategies • Data manipulation ■ Structure, Transform, edit • Selecting: ■ details on demand ■ Extract, Manipulate • Linking: ■ Brushing+linking • Filtering: ■ Dynamic Queries • Rearranging: ■ Sorting, ordering • Remapping: ■ Changing the visual mapping • Navigation: ■ Overview navigation: z+p, o+d, f+c, physical nav ■ 3 D navigate
Filtering/Limiting • Fundamental interactive operation in infovis is changing the set of data cases being presented ■ Focusing ■ Narrowing/widening ■ Suppression (remove irrelevant items from attention) • Iterative interactive queries/ dynamic queries
Sensitivity ? ? ? Home. Finder (Williamson and Shneiderman, 1992)
Visualizing Relations
When is Graph Visualization Applicable? • Ask the question: is there an inherent relation among the data elements to be visualized? ■ If YES – then the data can be represented by nodes of a graph, with edges representing the relations. ■ If NO – then the data elements are “unstructured” and goal is to use visualization to analyze and discover relationships among data. Source: Herman, Graph Visualization and Navigation in Information Visualization: a Survey
Graph and Tree Data Structures • Graphs: ■ Representations of structured, connected data ■ Consist of a set of nodes (data) and a set of edges (relations) ■ Edges can be directed or undirected • Trees: ■ Graphs with a specific structure • connected graph with n-1 edges ■ Representations of data with natural hierarchy ■ Nodes are either parents or children
Traditional Graph Drawing poly-line graphs (includes bends) orthogonal drawing planar, straightline drawing upward drawing of DAGs
Aesthetic constraints • • • Minimize link crossings Minimize link lengths Minimize link bends Maximize symmetries Minimize link crossings Mathematically difficult to do everything • Often unsuitable for interactive visualisation • • • Approximation algorithms very complex Unless you only need to compute layout once Precompute layout, or compute once at the beginning of an application then support interaction
(Some) Layout Approaches • • Tree-ify the graph - then use tree layout Hierarchical graph layout Radial graph layout Optimization-based techniques ■ Includes spring-embedding / force-directed layout • • Adjacency matrices Structurally-independent layout On-demand revealing of subgraphs Distortion-based views ■ Hyperbolic browser • (this list is not meant to be exhaustive)
Optimization-based layout • Specify constraints for layout ■ Series of mathematical equations ■ Hand to “solver” which tries to optimize the constraints • Examples ■ Minimize edge crossings, line bends, etc ■ Multi-dimensional scaling (preserve multi-dim distance) ■ Force-directed placement (use physics metaphor) • Benefits ■ General applicability ■ Often customizable by adding new constraints • Drawbacks ■ Approximate constraint satisfaction ■ Running time; “organic” look not always desired
Hyperbolic Layout • • • Root mapped at center Multiple generations of children mapped out towards edge of circle Drawing of nodes cuts off when less than one pixel
Presentation IAT 355 Introduction
Approaches • • • Viewport/Window (Panning and Zooming) Overview + Detail Distortion-based Views Focus + Context Time Based Views ■ RSVP
Vewport Scrolling hides most of a document Figure 4. 2
Panning is the smooth movement of a viewing frame over a 2 D image Figure 4. 34
Zooming • Geometric Zooming ■ Get close in to see information in more detail ■ Example: Google earth zooming in • Intelligent Zooming ■ Show semantically relevant information out of proportion ■ Example: speed-dependent zooming, Igarishi & Hinkley • Semantic Zooming ■ Zooming can be conceptual as opposed to simply reducing pixels ■ Example tool: Pad++ and Piccolo projects • http: //hcil. cs. umd. edu/video/1998_pad. mpg
Focus + Context Methods • • Filtering Overview+Detail Highlighting Distortion
Focus + context. Miniatures of pages of a pdf document provide useful context while attention is paid to detail of one page Figure 4. 3
Overview + Detail: “you are here” K. Hornbaek et al. , Navigation patterns and Usability of Zoomable User Interfaces with and without an Overview, ACM TOCHI, 9(4), December 2002 .
Arc. Trees – Interaction – Focus + Context Expanded relations between Ch 15 & 16 Hidden relations within Ch(16) of Glyph
Distortion-based Views • Distort an image of a large amount of information so that it can fit in screen. ■ Allow the user to examine a local area in detail; ■ At the same time, present a global view of the information space; • Provide navigation mechanism. • Co-existence of local details with global context at reduced magnification. • A focus region to display detailed information. • De-magnified view of the peripheral areas is presented around the focus area. Slide adapted from Fengdong Du
Distortion-based Techniques • • • Bifocal Display Polyfocal Display Perspective Wall Fisheye View Graphical Fisheye View Slide adapted from Fengdong Du
Figure 4. 8 Metaphor illustrating the principle of the Bifocal Display
Perspective Wall • Similar to Bifocal, except demagnifies at increasing rate, while Bifocal is constant • Visualizes linear information such as timeline • Adds 3 D but uses excess real estate on screen Slide adapted from Hornung & Zagreus
Fisheye Terminology • • Focal point Distance from focus Level of detail Degree of interest function Feb 28, 2011 IAT 355 141
http: //www. cs. umd. edu/hcil/fisheyemenu-demo. shtml
Slide adapted from Fengdong Du Image from Sarkar & Brown ‘ 92
- Slides: 143