Stronger Research Reporting Using Visuals

transcript

Applying Visual Design for Better Research – VCU Workshop

5th October, 2011

We live in a time of unprecedented

Information Overload

“ The highest-paid person in the first half of the next century

will be the ‘storyteller.’ ”

Rolf Jensen, 1996

As Story-tellers, we learn..

To write for the reader, not for yourself

A story needs a logical flow

To have a point of view

Only to report data that is vital to

telling the story

How can visuals help in storytelling?

Attention

Comprehension

Complexity

Understanding

Retention

Aesthetics

Timing

Emotion

The eyes are drawn like a magnet to images.

Graphics reduce time required to explain.

What’s wrong with wanting it to look good?

Pictures do a far better job of communicating emotion, and emotion does a far better job of inspiring action.

Presence of illustrations significantly improves retention.

Best way to summarise / represent complexity.

Less cognitive processing required, especially if image is familiar.

Can reveal patterns and relationships that would otherwise be hard to interpret or spot

Types of Visuals

Graphs

Illustrations

Data Viz

“Best 100 non-fiction books of the twentieth

century”

- amazon.com

Graphs

“When a graph is made, quantitative and categorical information is encoded by a display method. Then the

information is visually decoded. This visual perception is a vital link. No matter how clever the choice of the

information, and no matter how technologically impressive the encoding, a visualization fails if the decoding fails.”

(William S. Cleveland, The Elements of Graphing Data, Hobart Press, 1994, p. 1)

To 3D or not to 3D?

Series 1

Series 30

Series 1Series 2Series 3

To 3D or not to 3D?

Series 1

To 3D or not to 3D?

Series 10

Series 1

To 3D or not to 3D?

Series 1

Category1

Category2

Category3

Category4

Series 1

To 3D or not to 3D?

Series 1

To 3D or not to 3D?

Series 1

To 3D or not to 3D?

Series 1

Category 1 Category 2 Category 3 Category 4

Series 1

Series 2

Series 3

To 3D or not to 3D?

1st Qtr

2nd Qtr

3rd Qtr

4th Qtr

Losing Perspective

1st Qtr

2nd Qtr

3rd Qtr

4th Qtr

Losing Perspective

Areas, Volumes and Magnitudes

Ratio of size from Cat 1 to 2 is 1:2 BUT ratio or shape area is 1:4

Category1 Category

2Category

3Category

Lie factor = size of effect shown in graphic / size of effect in data

Series 3

Series 2

Series 1

Who ate all the Pies?

Sales 1st Qtr

2nd Qtr

3rd Qtr

4th Qtr

Sales 1st Qtr

2nd Qtr

3rd Qtr

4th Qtr

Sales 1st Qtr

2nd Qtr

3rd Qtr

4th Qtr

We make angle judgments when we read a pie chart, but we don’t judge angles very well. These judgments

are biased; we underestimate acute angles (angles less than 90°) and overestimate obtuse angles

(angles greater than 90°). (Naomi Robbins, Creating More Effective Graphs, Wiley, 2005, p. 49)

8% Sales 1st Qtr

2nd Qtr

3rd Qtr

4th Qtr

Q1 Q2 Q3 Q4

Apples

Oranges

Bananas

Grapefruit

Hollands and Spence found that trends are best analyzed with line graphs than with a series of pie charts. When estimating trends with line graphs, people can use a slope estimation procedure; with pie charts, they must perform multiple size discriminations between pie slices.

Hollands JG, Spence I. Judgments of change and proportion in graphical perception. Hum Factors 1992;34:313-34.

Chart Junk & Data Ink

$0.00 $1.00 $2.00 $3.00 $4.00 $5.00 $6.00

Category 1

Category 2

Category 3

Category 4

Gillan and Richman found that participants had faster response times and were more accurate when the data-ink ratio was high than when it was low. In addition, integrated tasks (e.g., global comparisons or synthesis judgments) appear to be more affected by the data-ink ratio than are focused tasks (e.g., selecting the value of a data point).

Gillan DJ, Richman EH. Minimalism and the syntax of graphs. Hum Factors 1994;36:619-44

Lipkus I M , Hollands J G J Natl Cancer Inst Monogr 1999;1999:149-163, Oxford University Press

Recap…

Data Integrity – avoid:

Data Clarity – avoid:

1. 3 dimensional treatments

2. Tricks of perspective

4. Too many pies

3. Lie-factors of area or volume

5. Unnecessary clutter

6. A low data-to-ink ratio

Tufte’s 5 principles of GOOD information design

5. Completely integrate words, numbers and images

3. Show multivariate data (more than 2 dimensions)

1. Enforce visual comparisons between groups

4. Content driven—all about explaining the data

2. Show or suggest causality

1. Enforce comparison

In other words, we must always ask the question, “compared to what?”.

Fortunately, visual comparison is faster and easier than mathematical or conceptual comparison:

“visualization made it possible to see the effects

of design changes on the pressure distribution

of an airplane wing, for example. The same thing

could be done with number crunching in theory,

but it was a lot more immediate and obvious where

things went wrong when the model was actually

shown as an image”

- Robert Kosara, http://stat-computing.org/newsletter/issues/scgn-22-1.pdf

simplified version

139 thousand tonnes of carbon dioxide would fill a sphere 521 metres across. To most Londoners, '139 thousand tonnes of carbon dioxide' is not a very meaningful quantity. Illustrating it in the context of London landmarks allows viewers to make it meaningful for themselves.

London’s Daily Greenhouse Gas Contribution 1. Enforce comparison

simplified version

1980’s weather is compared against ‘normal’ weather averages allowing you to immediately spot points of difference.

New York Weather for 1980 1. Enforce comparison

2. Suggest Causality

Without an indication of cause, you can be left wondering what the point is. i.e. if you show a trend, it begs the question, why is this happening?

http://youtu.be/pLqjQ55tz-U

2. Suggest Causality

simplified version

The world we seek to understand is multivariate.

The more variables, the more opportunities we have to see relationships and patterns

3. Show Multivariate Data

simplified version

3 Dimensions:- - Temperature - Precipitation - Humidity

New York Weather for 1980

oil consumption (Y axis) by year (X axis) and region (stacked area)

Increase in oil consumption

“Small Multiples” Also called Trellis / Lattice / Grid / Panel Chart

oil consumption (Y axis) by year (X axis) and region (stacked area)

Increase in oil consumption

Platforms, by type of usage, by volume

How BI Customers Use their Platforms

Platforms, by type of usage, by volume

How BI Customers Use their Platforms

Size of the circle is the amount of approval of

the premier/PM

Colour of Circle indicates vote difference

• Dark green = 15+ vote lead,

• Light green is 5-14 lead,

• White = +/- 5% lead/trail,

• Red= 5-14 trail & dark red (no example here) is

trail by 15 or more

Canadians think it time for a change of government, if they don’t see the

Government as being on the right track. And their vote intentions tend to

reflect that.

PEI NS

Time for Change

If there are elements that don’t serve the purpose of explaining the data, they are probably chart junk.

4. Content-Driven

There is nothing on here that is irrelevant New York Weather for 1980 4. Content-Driven

Aim for the viewer to be able to take in the whole picture in one glance, so avoid separate, complex legends which need to be continually referenced to make sense of the data

4. Fully integrate words, numbers and images

simplified version

Key annotations are present right within the chart New York Weather for 1980 4. Fully integrate words, numbers and images

Key annotations are present right within the chart

Distinct Segments driven by exposure interactions and psychographic engagement 4. Fully integrate words, numbers and images

Napoleon’s March on Moscow illustrates the principles

Enforce visual

comparisons —

the width of the

tan and black

lines gives you an

immediate

comparison of the

size of

Napoleon’s army

at different times

during the march

Show causality — the map shows how the

temperature and river crossings defeated Napoleon.

multivariate

data —

Napoleon’s

March shows

six: army size,

location (in 2

dimensions),

direction, time,

temperature

Completely

integrate

words,

numbers and images—in this

map, number

sit comfortably

with words and

the only legend

is a scale to

give a sense of

distance

The design

should be

content-driven —

Napoleon’s March

was designed as

an anti-war

poster…the

designer was

passionate about

the information

being presented.

The point of the

poster wasn’t the

design, it was the

information. simplified version

Quiz: Does this meet all of the criteria?

simplified version

Data Visualization

“Statistics journals rarely cover graphical methods… Outside of statistics, though, infographics and data visualization are more

important. Graphics give a sense of the size of big numbers, dramatize relations between variables, and convey the complexity of data and

functional relationships… sometimes to more efficiently portray masses of information that their audiences want to see in detail (as with sports scores, stock prices, and poll reports), sometimes to help tell a story (as

with annotated maps), and sometimes just for fun:.”

- Visualization, Graphics, and Statistics, Andrew Gelman and Antony Unwin, Statistical Computing & Graphics, July, 2010

Info-graphics

Data Visualization

Dashboards

Dynamic Data

Visualization

Info-graphics Summarize complex information using both decorative as

well as data-driven visual elements

Info-graphics

Dynamic Data Vizualisation

Uses motion or other interactive elements to allow the user

themselves to explore a dataset for insight

Dynamic Data Vizualisation

- Some tools becoming available

Many Eyes, (www.many-eyes.com)

Dashboards

Summarize key statistics into one page or panel of charts

Dashboards

Using Excel ‘Slicers’ for a Dynamic Dash

MY AWESOME DASHBOARD

Gross Profit

Sum of Total GP

Sum of Total sales

0 50000 100000 150000 200000

Salesperson 1

Salesperson 2

Salesperson 3

Salesperson 4

Salesperson 5

Total Sales

Sum of Total GP

Sum of Total sales

Jan-09

Feb-09

Mar-09

Apr-09

May-09

Jun-09

Jul-09

Product

Product A

Product B

Product C

Product D

Salesperson

Salesperson 1

Salesperson 2

Salesperson 3

Salesperson 4

Salesperson 5

Illustrations

“Ask yourself this: What information are you representing with the written word on a slide that you could replace with a photograph (or other appropriate image or graphic)?.. Images are powerful, efficient

and direct. Images can also be used very effectively as mnemonic devices to make messages more memorable. If people cannot listen and read at the same time, why do most PowerPoint slides contain far more

words that images? … It takes the realization that modern presentations with slides and other multimedia have more in common with cinema

(Images and narration) …than they do with written documents.”

- Presentation Zen, Garr Reynolds, 2008

Illustrations

Use of decorative, non-data driven images to add meaning

to your reporting.

Think like a designer: Simple, bold, colour-matched to your palette, Rule of 3rds

Use images along with bold words to make your headline points

For memorability or to emphasise a point pick an image that has an emotional appeal cute, comical, evocative

Source images from good quality, legal sources

Don’t be afraid to try!

But you don’t need to be one: a tonne of image manipulation tools right in PowerPoint.

Hzsdgldflgjfdkgsdfmgdfzgmfkgnfngzdmfng.djfngzdjfngngjnfgkldnfgzk;jngzjdfg.dfngzd/lfgKJBSDjfgbSLKDJFkjsdbfgKLJSDBklgSBDSLdjgbKSLjgbSLgbKSLJGBLDSFgbLKSJgbKSLJgbKLSJdgbLKSJgbKSLJgbLKJSgbklJSBDklggbLKSJgbkljSDBlgkJSgblkSJDBgklJSgbklSgbKLSJgblkSJDgblKSdjgbLKSgbKLSgbDSLJGlbKSJgblKSJgbLKSJgbKLSgbLKSgb SLDjgBLKSGBKLgbLKSJgbKLJSgbKLJSDgkjBSkljgSLDgjHBSDkljgBSDLGKSJhgbLKSgbKSLGBlSgbSgbSLkjgbSLKJGBSDLkjgbSgSLgbSLFGBDLkjgSBLGKjDSBlgkSBDkljgBSDkljgJBSDklgSkjghbSdlgKJSgbKLSJBGlkJSBkgBSDLKJGBLkdjgbSLKGJBKSLJGhbkdsljg

That’s all folks! Questions? Contact me: Laura Davies, SVP Laura.davies@visioncritical.com

Stronger Research Reporting Using Visuals

Business