Beat Signer - Department of Computer Science - [email protected] 2 May 4, 2023 Interaction ▪ Interactivity is necessary for vis tools handling complexity ▪ limitations of people and displays make it impossible to show a large dataset at once ▪ change level of details ▪ show different aspects of a dataset ▪ different representations and summaries of data ▪ different presentations of data
Beat Signer - Department of Computer Science - [email protected] 3 May 4, 2023 Handling Visual Complexity ▪ There a 5 major approaches for dealing with visual complexity ▪ deriving new data to be shown in a view ▪ view manipulation (single view) ▪ reduction (items or attributes) ▪ faceting into multiple views ▪ embed: focus+context (single view)
Beat Signer - Department of Computer Science - [email protected] 6 May 4, 2023 Facet Into Multiple Views … ▪ Faceting information into multiple views to deal with visual complexity ▪ Juxtapose views side by side ▪ coordination of views to create linked views (multiple views) - sharing of encoding - sharing of data - synchronisation of navigation ▪ arrangement of views ▪ when to show views ▪ Superimpose views as layers ▪ partitioning of elements between layers ▪ number of layers ▪ …
Beat Signer - Department of Computer Science - [email protected] 7 May 4, 2023 Sharing of Encoding ▪ Shared encoding views ▪ all channels handled the same for an identical visual encoding ▪ Multiform views (multiple views) ▪ some aspects of the visual encoding different between views ▪ each view might show a subset of attributes to avoid visual clutter ▪ Linked highlighting (brushing) ▪ items interactively selected in one view are highlighted with the same colour in all other views - see how region that is continous in one view is distributed in another
Beat Signer - Department of Computer Science - [email protected] 8 May 4, 2023 Exploratory Data Visualizer Example ▪ Linked highlighting (brushing) between views
Beat Signer - Department of Computer Science - [email protected] 9 May 4, 2023 Sharing of Data ▪ Shared data ▪ each view shows all the data ▪ Overview-detail (subset) ▪ one view shows entire dataset (overview) and another view shows detailed information about a selected subset of the data ▪ different possible combinations of views - subset data and shared encoding (e.g. bird's-eye map) - multiform views (e.g. details-on-demand view) ▪ Small multiples ▪ multiple views with same visual encoding but different partitions of the data - often aligned in matrix to support comparison between datasets - limited screen estate to show many views next to each other
Beat Signer - Department of Computer Science - [email protected] 10 May 4, 2023 Bird's-Eye Maps Example Bird's-Eye Maps What(Data) Geographic. How(Encode) Use given geographic information. How(Facet) Partition into two views with same encoding, overview-detail. How (Reduce) Navigate.
Beat Signer - Department of Computer Science - [email protected] 12 May 4, 2023 Partition Into Views ▪ Divide data up between views based on attributes ▪ partitioning attribute typically a categorical variable with limited number of values (levels) ▪ List alignment or matrix alignment ▪ Recursive subdivision ▪ partitioning used in an exploratory way ▪ user can reconfigure the display to see different choices of partitioning variables
Beat Signer - Department of Computer Science - [email protected] 13 May 4, 2023 Partitioning Bar Charts Example Single bar chart with grouped bars Aligned small-multiple bar chart views
Beat Signer - Department of Computer Science - [email protected] 18 May 4, 2023 Superimpose Views as Layers ▪ Design choices when superimposing multiple layers ▪ number of layers - often just two layers (e.g. background and foreground layer) ▪ distinction of layers - use different and non-overlapping range of the visual channels ▪ static vs. dynamic layers ▪ partitioning of items into layers
Beat Signer - Department of Computer Science - [email protected] 19 May 4, 2023 Cartographic Layering Example Cartographic Layering What(Data) Geographic. How(Encode) Area marks for regions (water, parks, other land), line marks for roads, categorical colourmap. How(Facet) Superimpose: static layers distinguished with colour saturation, colour luminance, and size channels.
Beat Signer - Department of Computer Science - [email protected] 20 May 4, 2023 Superimposed Line Charts Example Superimposed Line Charts What(Data) Multidimensional table: one ordered key attribute (time), one categorical key attribute (machine), one quantitative value attribute (CPU utilisation). How(Encode) Line charts, coloured by machine attribute. How(Facet) Superimpose: static layers, distinguished with colour. Scale Ordered key attribute: hundreds. Categorical key attribute: one dozen.
Beat Signer - Department of Computer Science - [email protected] 24 May 4, 2023 Embed: Focus+Context ▪ Single view embedding detailed information (focus) and overview information (context) ▪ avoid potential for disorientation that might result from navigation techniques such as geometric zooming ▪ Focus changes dynamically based on user interaction
Beat Signer - Department of Computer Science - [email protected] 25 May 4, 2023 Elide and Superimpose Data ▪ Some items are completely omitted from the view while others are summarised (context) ▪ dynamic aggregation for context ▪ only focus items shown in detail ▪ Superimposed layers can also be used for focus and context ▪ e.g. toolglass and magic lens example
Beat Signer - Department of Computer Science - [email protected] 27 May 4, 2023 Toolglass and Magic Lenses Example Toolglass and Magic Lenses What(Data) Spatial, quantitative curvature attribute across surface. How(Encode) Use given, colour by curvature. How(Reduce) Embed: superimpose.
Beat Signer - Department of Computer Science - [email protected] 28 May 4, 2023 Distortion ▪ Geometric distortion of contextual regions to make more space for details in focus region ▪ Design choices ▪ number of focus regions ▪ shape of focus region (e.g. radial or rectangular) ▪ global or local focus region ▪ interaction metaphor (e.g. movable lenses) ▪ Costs and benefits ▪ good for topological network structures ▪ length judgements severely impaired ▪ users might not be aware of distortion (potential misunderstanding)
Beat Signer - Department of Computer Science - [email protected] 29 May 4, 2023 Fisheye Lens Example Fisheye Lens What(Data) Any data. How(Encode) Any layout. How(Reduce) Embed: distort with fisheye; single focus, local radial region, moveable lens interaction metaphor.
Beat Signer - Department of Computer Science - [email protected] 30 May 4, 2023 Hyperbolic Geometry Example Hyperbolic Geometry What(Data) Tree or network. How(Encode) Hyperbolic layout. How(Reduce) Embed: distort by projecting from hyperbolic to Euclidean space; single global radial focus; hyperbolic translation interaction metaphor.
Beat Signer - Department of Computer Science - [email protected] 31 May 4, 2023 Nonlinear Magnification Fields Example Nonlinear Magnification Fields What(Data) Any data. How(Encode) Any layout. How(Reduce) Embed: distort with magnification fields; multiple foci, local arbitrary regions, lens or stretch or data-driven interaction metaphors.
Beat Signer - Department of Computer Science - [email protected] 32 May 4, 2023 Graph Exploration Example Fisheye lens Magnifying lens Neighbourhood highlighting with layering Neighbourhood highlighting with both layering and Bring and Go interaction
Beat Signer - Department of Computer Science - [email protected] 34 May 4, 2023 Further Reading ▪ This lecture is mainly based on the book Visualization Analysis & Design ▪ chapter 12 - Facet Into Multiple Views ▪ chapter 14 - Embed: Focus+Context
Beat Signer - Department of Computer Science - [email protected] 35 May 4, 2023 References ▪ Visualization Analysis & Design, Tamara Munzner, Taylor & Francis Inc, (Har/Psc edition), May, November 2014, ISBN-13: 978-1466508910