DSCI 532: Data Visualization II
Time: 11am-12:20pm Mon/Wed, Mar 20 - Apr 12 2017
Location: ORCH 3058
Labs: Mon 2-4pm, ESB 1042
Quizzes: Mon Apr 3 2-2:30pm, Thu Apr 20 2-2:30pm
Office hours: Mon 5:30-6:30 Mar 20 - Apr 10; Tue 2-3pm Apr 18 ICICS/CS X661
Slack channel: https://ubc-mds.slack.com/messages/531_viz-2
Instructor: Tamara Munzner; github @munzner, slack @tmm
Teaching fellow: Vincenzo Coia; github/slack @vcoia
TA: Sam Hinshaw; github @hinshaws, slack @samhinshaw
This module continues with analysis, design, and implementation considerations for creating interactive visualizations including multiple linked views, in contrast to the focus of DSCI 531 (Data Visualization I) on the creation of single static figures.
Schedule
Lect | Date | Day | Topic |
---|---|---|---|
1 | 2017-03-20 | Mon | Interaction design: reactivity, responsiveness, and selection. |
2 | 2017-03-22 | Wed | Interactive navigation: panning, zooming, and other changes of viewpoint. |
3 | 2017-03-27 | Mon | Design considerations for multiple views: juxtaposing and coordinating views. |
4 | 2017-03-29 | Wed | Design considerations for multiple views: partitioning and layering. |
5 | 2017-04-03 | Mon | Data reduction: filtering and aggregation of items. |
6 | 2017-04-05 | Wed | Data reduction: filtering and aggregation of attributes. |
7 | 2017-04-10 | Mon | Usability, user testing, and validation. |
8 | 2017-04-12 | Wed | Case studies of appropriate design choices and their trade-offs. |
Labs
Week/Assignment | Due | Lab topic |
---|---|---|
1 html, 1 md | Mon 3/27/17 9am | Interactive/reactive visualizations with shiny
|
2 html, 2 md | Mon 4/3/17 9am | Interactive multiple linked views with shiny and ggmaps , multiple views with ggplot2
|
3 html, 3 md | Mon 4/10/17 9am | Interactive data filtering and aggregation with shiny
|
4 html, 4 md | Tue 4/18/17 9am | Peer review of usability, iterative refinement of previous software |
Quizzes
Time | Date | Location | |
---|---|---|---|
1 html, 1 md | 14:00 - 14:30 | 2017-04-03 | ESB 1042 |
2 html, 2 md | 14:00 - 14:30 | 2017-04-20 | ESB 1042 |
Solutions
Reference Material
-
Overall
- Munzner, Tamara. Visualization Analysis and Design, CRC Press, 2014. ebook link, Tamara's book page
-
Week 1: Lectures 1/2, Mon Mar 20 & Wed Mar 22. Manipulate & Interact.
- Slides: Lectures 1/2 1up PDF, 16up PDF
- Further reading: VAD Chapter 11
- Demos/Videos:
- Growth of a Nation
- Sortable Bar Chart
- DataStripes
- LineUp Demo
- video: Animated Transitions
- Stacked to Grouped Bars
- Hierarchical Bar Chart
- Collapsible Tree
- The Scrollytelling Scourge, Kosara
- How to Scroll, Bostock
- Scrollytelling: Bloomberg Warming
- Scrollytelling: NYT Oil
- Zoom to Bounding Box
- Zoomable Icicle
- video: Multilevel Matrix zoomin
- video: Multilevel Matrix zoomout
- video: Multilevel Matrix pan
- Every Map Projection
- Tissot Indicatrix
- Myriahedral Projections
- bonus: What Your Favorite Map Projection Says About You
-
Week 2: Lectures 3/4, Mon Mar 27 & Wed Mar 29. Facet into Multiple Views.
- Slides: Lectures 3/4 1up PDF, 16up PDF
- Further reading: VAD Chapter 12
- Demos/Videos:
- Pointers from discussion
- Followup reading after lab
- Effectiveness of Animation in Trend Visualization Robertson et al, IEEE TVCG 14(6):1325-32 (Proc InfoVis 2008).
-
Week 3: Lectures 5/6, Mon Apr 3 & Wed Apr 5. Reduce Items & Attributes
- Slides: Lectures 5/6 1up PDF, 16up PDF
- Further reading: VAD Chapter 13, VAD Chapter 14
- Demos/Videos:
- Multivariate Network Exploration scented histogram bisliders, compact
- Vismon scented histogram bisliders, detailed staged
- Crossfilter
- Buy/Rent Calculator crossfilter in NYTimes article
- Continuous Scatterplots
- Probing Projections
- Using t-SNE Effectively
- Fisheye Lens D3
- TreeJuxtaposer
- Bring & Go
-
Pointers from discussion
Week 4: Lectures 7/8, Mon Apr 10 & Wed Apr 12. Usability & Case Studies
- Slides: Lectures 7/8 1up PDF, 16up PDF
- Further reading:
- Book: VAD Chapter 15
- Book Case Studies
- Graph-Theoretic Scagnostics Leland Wilkinson, Anushka Anand, and Robert Grossman. Proc InfoVis 05.
- Scagnostics Distributions Leland Wilkinson and Graham Wills. Journal of Computational and Graphical Statistics (JCGS) 17:2 (2008), 473-491.
- VisDB: Database Exploration using Multidimensional Visualization Daniel A. Keim and Hans-Peter Kriegel, IEEE CG&A, 1994
- Interactively Exploring Hierarchical Clustering Results Jinwook Seo and Ben Shneiderman. IEEE Computer 35:7 (2002), 80-86.
- A Rank-by-Feature Framework for Interactive Exploration of Multidimensional Data Jinwook Seo and Ben Shneiderman. Information Visualization 4:2 (2005), 96-113.
- Visual Exploration of Multivariate Graphs Martin Wattenberg, Proc. CHI 2006.
- InterRing: An Interactive Tool for Visually Navigating and Manipulating Hierarchical Structures Jing Yang, Matthew O. Ward, and Elke A. Rundensteiner. Proc InfoVis 2002, 77-84.
- Constellation: A Visualization Tool For Linguistic Queries from MindNet Tamara Munzner, Francois Guimbretiere, and George Robertson. Proc. InfoVis 1999, p 132-135. Constellation video
- Constellation: Linguistic Semantic Networks Tamara Munzner, Interactive Visualization of Large Graphs and Networks (PhD thesis) Chapter 5, Stanford University, 2000, pp 87-122
- UX/HCI:
- 7 Step Guide to Guerrilla Usability Testing
- The Art of Guerrilla Usability Testing
- Discount Usability: 20 Years
- Interaction Design: Beyond Human-Computer Interaction Comprehensive HCI textbook, recently updated.
- About Face: The Essentials of Interaction Design Interaction design from a practitioner point of view, quite complete, recently updated.
- Task-Centered User Interface Design Online book from the 90s, now completely free.
- Designing with the Mind in Mind Highly readable distillation of many cognitive principles relevant for UX (user experience) designers.
- Theory: An Algebraic Process for Visualization Design
- Design/Redesign Process In Depth:
- Dear Data Two visualization designers gather data and create visualizations sent to each other on a hand-drawn postcard, one specific topic each week for a full year; each post includes reflections about data gathering and the design process.
- Data Sketches Two visualization designers each do a substantial project each month on the same theme, with extensive notes on process and design choices and computational methods and tests.
- Makeover Mondays (Tableau) Each week dozens to hundreds of people do a redesign of a chosen chart in Tableau, intended to be a quick thing doable in just a few hours; blog post includes critique/analysis of several of the standouts.
- Theory Practice Exercises from CS Visualization course
- Tableau tutorial exercises from Data Vis For Journalists course
- Resources
- Visualization books
- The Functional Art & The Truthful Art by Alberto Cairo Data journalism point of view, Cairo is a journalist who was a practitioner and is now a professor, highly readable.
- Data Visualization: A Handbook for Data Driven Design by Andy Kirk Visualization freelancer/consultant point of view, Kirk runs many short tutorials for practioners; very readable advice on the entire process of a visualization project.
- Design for Information by Isabel Meirelles Designer point of view; not just a coffee table book, includes a substantial discussion.
- Visual Thinking for Design by Colin Ware Cognitive psychology researcher point of view intended for a somewhat broad audience, a shorter read than his complete and research-oriented book Information Visualization: Perception for Design
-
Stephen Few books Business intelligence meets visualization precision/accuracy purist point of view, operationalizeable and readable.
- [Information Dashboard Design: Displaying Data for At-A-Glance Monitoring]
- [Show Me The Numbers: Designing Graphs and Tables to Enlighten]
- [Now You See It: Simple Visualization Techniques for Quantitative Analysis.
- The Elements of Graphing Data by Bill Cleveland, Visualizing Data by Cleveland Statistician point of view, thorough and comprehensive but not a light read.
- Creating More Effective Graphs by Naomi Robbins Robbins has distilled many of the principles covered by Cleveland into a very accessible and quick to read book - very operationalizable advice on how exactly to create and fine-tune statistical graphics/charts. Companion R site by Jenny Bryan and Joanna Zhao
- A Field Guide to Digital Color by Maureen Stone Next step for going deeper into color science, good balance between technical precision and readability.
- Pointers from discussion
Learning Outcomes
By the end of the course, students are expected to be able to:
- Analyze interactive visualizations in terms of approaches to handling complexity: dynamic change over time, partitioning into multiple views, and data reduction within a single view (in addition to the derivation of new data, as covered in Viz-1 module).
- Design new interactive visualizations for complex datasets.
- Implement interactive visualizations using existing toolkits and libraries.
- Explain the trade-offs of using animation vs juxtaposed views vs derived data.
- Explain and justify methods to validate visualization design effectiveness including computational benchmarks, field studies on deployed software, and qualitative discussion of visual results.
Lecture Learning Objectives
- Interaction design: reactivity, responsiveness, and selection.
By the end of the lecture, students are expected to be able to:
- Explain costs and benefits of interactively changing views.
- Reason about the design choices involved in selection and highlighting.
- Characterize human response to latency into categories relevant to the design of responsive software.
- Interactive navigation: panning, zooming, and other changes of viewpoint.
By the end of the lecture, students are expected to be able to:
- Explain the difference between geometric and semantic zooming and reason about when each is appropriate.
- Explain the tradeoffs between constrained and unconstrained navigation.
- Explain the tradeoffs between slicing, cutting, and projection.
- Design considerations for multiple views: juxtaposing and coordinating views.
By the end of the lecture, students are expected to be able to:
- Analyze the use of multiple views in terms of shared versus different visual encoding and data shown.
- Discuss the tradeoffs between a single view that changes over time and multiple linked views.
- Explain the costs and benefits of different view coordination strategies.
- Design considerations for multiple views: partitioning between views and visual layering.
By the end of the lecture, students are expected to be able to:
- Reason about the order in which to select attributes to partition a dataset into multiple views according to the target user's task.
- Analyze the scalability of superimposed layers for static layering
- Analyze the scalability of superimposed layers for dynamic layering
- Data reduction: filtering and aggregation of items
By the end of the lecture, students are expected to be able to:
- Discuss the tradeoffs of filtering vs aggregation for reducing data.
- Analyze static aggregation techniques in terms of what data was derived to support them
- Analyze dynamic aggregation techniques in terms of derived data and interactive selection.
- Data reduction: filtering and aggregation of attributes
By the end of the lecture, students are expected to be able to:
- Relate dimensionality reduction techniques to attribute aggregation goals
- Analyze complex combinations of filtering and aggregation.
- Usability and user testing. By the end of the lecture, students are expected to be able to:
- Conduct basic usability testing.
- Reason about the costs and benefits of different forms of validation and user testing to make appropriate choices given time constraints and quality targets.
- Case studies of appropriate design choices and their trade-offs. By the end of the lecture, students are expected to be able to:
- Analyze an existing visualization interface according to all previously discussed criteria
- Relate the scalability of visualization design choices to the four major strategies for handling visual complexity: deriving new data, changing the view over time, partitioning into multiple views, and reducing the amount of data shown in a single view.
- Propose a new visualization design appropriate for a specific data and task combination.