# What are the risks that can occur when working with data visualizations?

The American statistician Edward tufte* has done an analysis of data visualizations that he did not succeed.

* By some awarded as the “Da Vinci of Data visualization” and the “Galileo of graphs.”

Challenger problems

As a best-known example, Tufte gives the data visualizations that reflect the problems surrounding the O-rings of the space feather Challenger .

Morton Thiokol’s depiction of O-ring damage during the Challenger accident.

According to him, the Challenger would not be injured if the data visualizations were more brightly shaped, which would provide the problems.

In other words: better data visualization could avoid a great accident.

To demonstrate superior data visualization, Mr. Tufte has shown the data in a scatter chart, in which he clearly indicates the correlation between the temperatures of the rockets and the potential defects the O-rings.

Edward Tufte’s depiction of O-ring damage and proportional relationship with temperature

However, although his diagram appears to be visually brighter and more readable, it basically relies on a misinterpretation of the data; namely that there is no proportional correlation between temperature and the defects that are going to occur.

Tufte has misunderstood the importance of the temperature, which gave the data a new context, with the result that you cannot draw the right conclusion based on the scatter chart.

In addition, the original graphs were clear -although technically and for the layman, it was not easy to read -and the problem was also known.It has been decided to let the Challenger mission go through, among other things because they were under pressure.

This was not an example of a situation where a data designer could have saved lives -certainly not in the air.

Risks of data visualizations

Nonetheless, Tufte unknowingly shows the risks of working with data visualizations.These risks concern (among others):

• Data visualization is intrinsically reducing.
• Data visualization seems to reflect facts, but is nonetheless part of a (subjective) narrative.
• People tend to believe what they want, and allow themselves to be easily guided by visualizations.
• Incomplete data is worse than incorrect data (incorrect data is identifiable; incomplete data often not).
• Correlation does not imply a cause yet.
• If the data visualization is confused, the rationale behind it is guaranteed to be confused.
• The context is important for both encoding and decoding the data as well as visualization.

Data needs the right context to function as information and to be read correctly.If the context is not correct, there is no relationship with the information, with the result that no (justified) knowledge can be made.

Na:

1. Data -Unprocessed, disorganized
2. Information -Edited, contextualized
3. Knowledge -Narrative