Designing Effective Data Visualizations

Pedro Cardoso-Leite

pedro.cardosoleite@uni.lu

Let’s start with a story…

The night before…

What’s the issue?

O-rings
might get damaged
when it’s too cold.

What happened?

What was the concern?
What data / information should they have shared?

Of the 13 charts circulated by Thiokol managers and engineers to the scattered teleconferees,

six contained no tabled data about either O-ring temperature, O-ring blow-by, or O-ring anomaly (these were primarily outlines of arguments being made by the Thiokol engineers).

Of the seven remaining charts containing data,

six of them included data on either launch temperatures or O-ring anomaly
but not both in relation to each other.

Only one chart relating O-ring temperature and O-ring damage was prepared by a Thiokol engineer.

Let’s focus on this one document.

What do you think about this chart?

One more story…

There are now more modern version of the John Snow data exploration

Let’s have a look at some numbers…

What can you tell me about these four datasets?
Are they the same?

What can you say about these four datasets?

Anscombe’s Quartet

Constructed in 1973 by statistician Francis Anscombe
Demonstrates the importance of graphing data
Shows the effect of outliers on statistical properties
Intended to counter the impression among statisticians that “numerical calculations are exact, but graphs are rough.”