In today’s technology driven world we are faced with ever-growing quantities of data. Time and time again we ask ourselves: How do we best present this data?
Edward Tufte offers some advice in the area of data visualisation.
Data Visualisation
Chartjunk. One of the terms that Tufte has coined is “chartjunk”. Chartjunk refers to the elements in a chart which do not tell the viewer anything new and, even worse, could potentially distract the viewer from the most important message that is to be conveyed. The following are examples of “chartjunk”:
- Heavy grid lines
- Unnecessary text
- Too much ornamentation (e.g. ornamental shading)
- Poor scale and too many dimensions
Data-Ink Ratio. Another concept that Tufte has invented is the data-ink ratio, that is, the proportion of ink used to show the data, versus the ink used to show the overall graphic. Minimalist charts would have a high data-ink ratio, whilst excessively decorated charts would a low data-ink ratio. The goal is to remove any unnecessary elements so that we increase the proportion of ink used for showing the actual data.
Data Density. A related concept is data density, which refers to the proportion of the total graph which is devoted to displaying data. The goal is to create graphs with high data density, whereby most of the graph is occupied by data, as opposed to redundant non-data elements.
Shrink Principle. According to this principle, it is often possible to shrink graphs without losing information.
Small multiples and Sparklines. There are numerous occasions when we have numerous data series, which would appear conglomerated if placed on a single chart. When we have data recorded over multiple axes (i.e highly dimensional data) we face similar problems, compounded by potential confusion faced by the viewer. Tufte’s solution is to break down this data and show it as “small multiples” whereby a chart may be broken down and displayed as a series of separate charts placed next to each other. One example where this can be used is for stock market data. Tufte’s sparklines, data-rich graphics, that fit on lines without breaking the flow of sentences, naturally flow from this.
What about data in presentations?
Tufte is a strong critic of the way PowerPoint styles disrupts and trivialises data. This criticismis evident in his essay “The Cognitive Style of PowerPoint” and was further stressed in his criticism of the use of PowerPoint by NASA engineers. So, what’s wrong with PowerPoint? The following are some of the issues identified:
- Low resolution, low content
- Deeply nested hierarchies
- Simplistic bullets that omit words and relationships
- Forced linear progression of ideas
- Data is broken up over slides, preventing meaningful comaprsion
Tufte’s recommendations to presenters who want to improve their presentation of data include providing handouts or short terports, especially in settings which require the communication of technical information.
Interesting links
- Data-Ink Ratio
- Edward Tufte on Columbia Evidence (2003)
- PowerPoint Does Rocket Science--and Better Techniques for Technical Reports
Post new comment