critique des statistiques

Désolé, mais cette section sera en anglais.

Voici un lien à un article nommé:Six Simple Techniques for Presenting Data: Hans Rosling (TED, 2006) 

Six Simple Techniques for Presenting Data: Hans Rosling (TED, 2006)

by Andrew Dlugan

Published: Jan 9th, 2008

Hans Rosling presented a fantastic talk at TED. The delivery was inspiring, the mood was electric, and it was all about statistics. Yes, statistics – a topic most often associated with dry and boring presentations.

Hans Rosling uses six simple techniques for presenting data which transform a run-of-the-mill presentation into a must-see presentation.

I encourage you to:

1.    Watch the video;

2.    Read the analysis in this speech critique; and

3.    Share your thoughts on this presentation.

Six Techniques for Presenting Data

Rosling employs GapMinder to display his statistics. This is a wonderful software tool for displaying data, but the real magic of this presentation lies in the techniques demonstrated by Rosling. These techniques are easy to do, but I’ve rarely (if ever) seen them all demonstrated so well in a single talk. The techniques are:

1.    Explain the data axes

2.    Highlight subsets of data

3.    Dig deeper to unwrap data

4.    Place labels close to data points

5.    Answer the “Why?” questions

6.    Complement data with energetic delivery

Let’s examine each one and compare this presentation to common approaches.

Technique #1: Explain the Data Axes

Common approach. Graphs are displayed with either no explanation of the axes, or a quick, obligatory “Here we see variableX versus variableY“.

As Hans demonstrates, don’t assume that your audience intuitively “gets it,” particularly when presenting statistical data.

Starting around 2:43, he devotes approximately ninety seconds to:

§  Explain what quantities are on each of the two axes (e.g. fertility rates versus life expectancy at birth);

§  Provide the background story as to why he chose these two quantities (“We vs Them = Western World vs Third World“);

§  Share his students’ prediction as to what the data will show.

Because of this careful preparation, the audience understands the context thoroughly. A very energetic description of the data follows while the time advances the “movie” for about 45 seconds.

The “instant reply” is a nice touch which fills the otherwise empty time during audience applause, although I suspect this was added in the post-production by the good folks at TED.

Technique #2: Highlight subsets of data

Common approach: Presenters attempt to explain complex data which they have studied for days, weeks, or months in just a few minutes. The audience grasps little.

Rosling recognizes the impossibility of explaining all of the data in detail. Instead, he carefully selects and explains subsets of the data.

§  Example #1: 1964-2003 United States and Vietnam [5:15 to 6:06] This is a clever choice as his (mostly American) audience will easily connect the early part of this period with that of the Vietnam War.

§  Example #2: 1960-2003. South Korea, Brazil, Uganda, United Arab Emirates [12:22 to 13:40]

Technique #3: Dig deeper to unwrap data

Common approach: Presenters restrict themselves to one level of data inspection. Deeper analysis is often only present in scientific journals.

Several times, Rosling displays first a high-level data view (e.g. one point for a country) and then digs deeper to lower-level view of the data (e.g. country quintiles).

§  Example #1: Income versus population
Compare the global curve [7:26] to the one broken down by geographic region [7:54]

§  Example #2: GDP per capita versus Child survival rate
Compare the Sub-Saharan Africa bubble [9:48] to the individual country bubbles [9:54]

§  Example #3: GDP per capita versus Child survival rate
Compare Uganda bubble [14:12] to quintile data points for Uganda [14:18]

Technique #4: Place labels close to data points

Common approach: Data legends and labels are often absent. The presenter assumes that the audience will follow their verbal cues. Or, when legends and labels are present, they are often presented far away from associated data points. This forces the audience to visually scan back and forth.

Throughout Rosling’s talk, data labels are presented right next to the data points. An example is shown here for the OECD data point [9:28].

Additionally, the appearance of these labels is synchronized well with the verbal component of his speech. In this way, the visual labels complement the audio.

Related to this, there are several instances where GapMinder shows a bubble “about to burst” a second or two before the data is expanded. This is a subtle touch, but an effective measure to draw the eye to the right spot on the screen. The Sub-Saharan Africa example shown is from 9:48. Others are at 10:33 and 10:40.

Technique 5: Answer the “Why?”questions

Common approach: Large data sets are presented, and the presenter often explains only the dominant trend or the one measure of most interest. The audience is left to wonder things like “Why is that data point there?” or “What caused that point to be low/high/odd?”

Obviously, no presenter can answer every question the audience might be thinking, but Rosling does a good job of anticipating these questions. He anticipated several “Why?” questions, and answered them on the spot.

For example:

§  Q: Why does the progress in Vietnam accelerate in the 1990’s? [5:43]
A: They give up communist planning and go for a market economy.

§  Q: Why is Mauritius so different from most of Sub-Saharan Africa? [9:54]
A: Mauritius was the first country to get rid of trade barriers. They could sell their sugar. They could sell their textiles.

§  Q: Why is China moving up and then to the right (when most countries are moving diagonally)? [11:52]
A: Mao Zedong bought health to China (up) and then he died. Deng Xiaoping then brought money to China (right).

Anticipating and answering the why questions achieves two goals:

1.    It allows you to satisfy the audience’s curiosity while also maintaining an energetic pace (rather than being interrupted by questions).

2.    It demonstrates your credibility and solid grasp of the subject.

Technique #6: Complement data with energetic delivery

Common approach: Statistical data is often presented in a dry, clinical manner. Perhaps the theory is that the audience should naturally be excited about data?

The most memorable technique displayed by Rosling is his energetic delivery. Examples are numerous, and include:

§  Highly energetic sequence as he narrates 1962-2003 fertility rates versus life expectancy [4:15 to 5:03]. This is the highlight of the presentation for me. The audience seems to agree, rewarding Rosling with 13 seconds of applause.

§  Spider-web shape with his hands to demonstrate how the bubbles burst [9:55]

§  Ghost-like acting to accompany “overlooking the United States, almost like a ghost” [18:11]

§  Approaching the screen numerous times to align his arms and body with the data

The thoughtful presentation of data makes this an understandable talk. Rosling’s energetic delivery makes it memorable.

Devoir: Choisissez une graphique, puis analysez-la en utilisant les 5 techniques.