Plotting Time Series Data

,

An infographic in the Raconteur supplement to the Times published on 11 November 2014 illustrates one of the classic pitfalls to avoid in plotting time series data with incomplete data sets.

The lower portion of the infographic shows an area chart that plots the following six numbers:

$bn 2013 2014 2018
3D Printers 0.7 1.3 5.4
Services & Materials 1.8 2.5 10.8

No numbers are available for 2015, 2016 and 2017.

The data appears like this in the infographic:

3D Printing

 

The scaling of the time axis for the area chart does not show the passage of time proportionately. The distance between 2014 and 2018 is only slightly greater than the distance beween 2013 and 2014, despite the fact that it covers a time range that is four times greater.  As a result, the growth over the period 2014 to 2018 is highly exaggerated over the single year of growth from 2013 to 2014.

The extent of this exaggeration can be seen by plotting the original data using a time axis that is in correct proportion.

Correct scaling

 

Compare the message in this chart to what happens when we mis-scale the axis by making the interval between 2014 and 2018 the same as that between 2013 and 2014.

Incorrect scaling

 

This is a classic misrepresentation of the data that, if repeated in a business context, could lead the reader to draw incorrect conclusions.

Leave a Reply