Part 1 – Overview and Study Hall
On August 6th, 2015 I took Professor Edward Tufte’s one day course entitled: Presenting Data and Information. The following three blog posts and videos are my attempts to summarize the five hours worth of material.
To do this effectively, I’m breaking the content into three distinct parts:
Part 1: Study Hall – I will cover the seven reading assignments and how they relate to the course.
Part 2: Tufte’s Course – I will share the topics covered and the notes I took throughout the day.
Part 3: Homework and Closing Thoughts – I will cover the homework assigned for the course and my closing thoughts.
Part 1: Study Hall
The seven reading assignments for the course are the following:
|Reading Number and Book||Page Numbers|
|1 – Envisioning Information||50-51 (End of Chapter 2)|
|2 – Visual Explanations||68-71 (End of Chapter 3)|
|3 – Visual Explanations||146-150 (End of Chapter 7)|
|4 – The Visual Display of Quantitative Information||13-31 (Beginning of Chapter 1)|
|5 – The Visual Display of Quantitative Information||46-51 (End of Chapter 1)|
|6 – The Visual Display of Quantitative Information||120-121 (End of Chapter 5)|
|7 – The Visual Display of Quantitative Information||177-183 (Beginning of Chapter 9)|
Now instead of reading just the recommended sections, I read the complete six chapters the readings spanned. My goal is to provide the best possible context as to how the readings fit within the course material by summarizing the key themes in each chapter.
Reading 1 – Envisioning Information – Chapter 2: Micro/Macro Readings
- To clarify your visual design, you may need to add detail
- Arbitrary boundaries are OK if they are statistically wise – Especially at the macro level
- The Vietnam Veterans Memorial does macro to micro incredibly well.
- Large datasets coupled with high density designs allows more viewer control to personalize the data for their own uses.
- Thin data should prompt suspicion.
- “Clutter and confusion are failures of design, not attributes of information” -pg51
- Simplicity is a preference, not an information display strategy.
Reading 2 – Visual Explanations – Chapter 3: Pictorial Instructions and Disinformation Design
- Many layers need to be depicted in communicating magic tricks visually. The audience view or what appears to be done. The magicians view or what’s actually done and audience misdirection.
- Diagrams have the ability to show activities which we wouldn’t be able to see the real world because of our fixed point of view.
- Good typeset can show voice, movement, vanishing and activity.
- Intentional disinformation design will divert scrutiny.
- The truth will not stand in the way of a good story.
- Don’t suppress context. Don’t prevent reflective analysis. Tell your audience what you are going to do ahead of time.
- What is the problem?
- Why the problem is important.
- What is the solution to the problem.
- Move from the particular to the general to the particular.
- Give everyone in the audience a handout packed with material related to your presentation.
- Have the handouts contain: names of the people at the meeting, research methods, pictures/diagrams, data tables
- Your integrity is reflected in your content. Master the details of your presentation.
Reading 3 – Visual Explanations – Chapter 7: Visual Confections: Juxtapositions from the Ocean of the Steams of Story
- “A confection is an assembly of many visual events…Confections illustrate an argument, show and enforce visual comparisons, combine the real and the imagined, an tell us yet another story.” -pg121
- Confections may use symbolism, compartments, callouts, poetry and metaphors.
- Confections let the information become an interface the viewer can quickly sort through.
- Don’t march your viewer through a decision tree with tiny irritating steps which reveal the software logic.
- A good confection makes reading, seeing and thinking identical. They are gifts in understanding.
Reading 4 – The Visual Display of Quantitative Information – Chapter 1: Graphical Excellence
- Graphical displays should communicate complex ideas with clarity, precision and efficiency.
- If you have simple linear changes, don’t waste time with data graphics, just use summary numbers.
- Graphics show the shape of the data in comparative context.
- The passage of time put into descriptive chronology is not a causal explanation.
- Well-designed graphics focus on substance, statistics, design, clarity, precisions and efficiency. They give the viewer a great number of ideas in a short amount of time.
- Always tell the truth about your data.
Reading 5 – The Visual Display of Quantitative Information – Chapter 5: Chartjunk: Vibrations, Grids, and Ducks
- Various reasons are invoked to justify decorations that don’t tell the viewer anything new. Artistic creativity, enliven the display, make the graphic appear more scientific and even to capture the essential “spirit of the data”.
- Eye-straining effects have no place in data design.
- “It is all right to decorate construction but never construct decoration.” – Pugin Augustus Welby Northmore (Duck effect)
- Don’t add fake perspectives to data graphics. Graphics either stand or fall based on their data content.
- No substance is really generated by chartjunk.
Reading 6 – The Visual Display of Quantitative Information – Chapter 9: Aesthetics and Technique in Data Graphical Design
- “Graphical elegance is often found in simplicity of design and complexity of data.” -pg177
- Combine two of these three basic structures for showing data: the sentence, the table or the graphic.
- Words need to tell the viewer how to read the design but integration of text and figures can be done well.
- Balance the relative proportions of the graphical elements. Don’t weight all lines equally.
- Follow the shape of the data towards the suggested shape of the graphic. Otherwise, use horizontal graphics that are about 50% wider than tall.
Part 2 – Tufte’s Course
Shortly after 10am, the entire room went dark. Two large projectors now displayed Stephen Malinowski’s – Music Animation Machine. Tufte added commentary on how we see the future on the right, the past on the left, and what’s happening now is lighting up. He added that how with a visualization such as this, we may actually hear the music better. From an audience perspective, I found this opening to produce a rather physically calming but mentally active state.
Tufte then showed us some preselected articles from the New York Times and ESPN which included data tables and confection displays. He also pointed out that users come to such websites and articles for a content experience, not a values experience. One should add voices of experts to lend credibility and stay away from cherry picking data. Showing the raw data and the large number-sets will only help your case.
For about 30 minutes Tufte covered how to give a great presentation and make meetings 20% shorter. Such material was also part of the study hall reading in his book Visual Explanations – pages: 68-71. His tips during the course were the following:
- Every meeting begins with a document.
- Every meeting begins with study hall.
- Do not share or let people read the material in advance.
- Maintain the meeting room as a sanctuary for the material.
- When fielding questions: take notes, don’t evade, don’t disparage your colleagues.
- Set an agenda ahead of time.
- When doing specifically scientific presentations with abstract focal points: focus on what the problem is, who cares about it and what is to be done about it.
Several practical advice points were also given around presentations:
- Did you get the story right?
- Do people believe you?
- The single biggest threat to credibility of a data presenter is cherry picking data.
- Be ready to become the target or anger if you are the agent provocateur with your presentation.
- Think the best of your audience.
- Bad presenters show things that are too good to be true.
- Bad presenters will not give the audience links to the source material.
- Bad presenters don’t master the details.
In addition to presenting, Tufte also gave several good points on how to be a good spectator during presentations:
- Stay focused on the content.
- View the presentation with an open mind but not a head so empty your brains fall out.
- If the absurdities start piling up, leave the room.
- Loot the presentation for useful material.
- Listen, see, think, learn.
Practical advice for thinking about data:
- Find what is relevant.
- Find out how the measures were collected. “You never learn more about a process than to look at how the data was collected.”
- People can’t keep their own score. If they do, the Principle of Lake Wobegon may take effect, where all the data is above average.
The day concluded with Tufte showing the audience a first printing of Euclid’s Elements from the late 15th century. The book contains several ways in which the reader can escape flatland and see objects as they would appear in the real world. This is the purpose of visualizations, to see better in the real-world where the data is actually happening.
Part 3 – Homework and Final Thoughts
Homework – The Parable of Google Flu
The reading assignment was more topical than illustrative of good data visualization. The article contained only two graphs which are simple, effective and lacking any chart junk. I added the area showing where the nonseasonal H1N1 influenza-like illness(ILI) were missed by Google Flu Trends.