07 Jul 2016

Tableau’s Scatter Plots: A Powerful Tool to Highlight Correlations

Nicolas Voirin

To demonstrate how to use Tableau's Scatter plots, I investigated the relation between all of the UK's previous Prime Ministers approval rating (until 2004) and the CAGR (Compound Annual Growth Rate) change in GDP and Unemployment during their term. Are British PMs' reputations connected to the economic impact they've had on the country?

Technically, Scatter Plots are not a challenging function to use in Tableau:

  • Choose X and Y Axes: drag the measures you wish to plot against each other in the Columns and Rows. This is effectively determining the X and Y axes of your Scatter chart.
  • Data Aggregation: Choose how you wish to aggregate your data, in this case, I wanted to aggregate by Prime Minister and see the economic factors during each term. See how I dragged Prime Minister to the Detail card.
  • Colour and Sizing: Scatters are powerful because you visualise more data using both Colour and Size. Here, I coloured each point by the Prime Minister score. 

Scatter Plot Marker 

Figure 1: Scatter Plot Marks Card

Scatter Plots can then be easily included in more sophisticated dashboards with highlight actions: in this case, I included the Prime Ministers' names, pictures, and a ranking of all the Prime Ministers until 2004 by the score they received in the Ipsos Mori survey data I used.

Interestingly enough, it seems that economic factors such as GDP and unemployment are far from being the only factors determining Prime Ministers' popularity. During Sir Douglas-Home's term, GDP grew by an astonishing 5.6% and unemployment was virtually demolished with a 35% reduction, yet the survey gave him a score of 3.3/10.

Nicolas Voirin

About the author

Nicolas Voirin - Graduate Analytics Consultant. I have started my career in the field of analytics because I have an interest in how data can be used to bring genuine value to organisations. With a background in Engineering and Management, I take pleasure in solving business problems and believe visualising data is a powerful tool to do this. I am curious about a range of topics but have previously enjoyed building data visualisations concerning politics and history.