24 May 2016

Superhero Analytics: How to format your Tableau Dashboard in Four Steps

Ryan Smith

To find out more about Concentra's bespoke Analytics Services and Solutions see Services.

Tableau is all about easy to produce, aesthetically pleasing visuals, but there can be a vast difference between visualisations created with or without formatting best practices - as the latter often lacks a certain finesse. This blog will show the four stages of building a Tableau dashboard and what to consider in each of them in order to deliver a polished visualisation.

My starting point is the dashboard below, showing information on superheroes in New York. We can all agree it could be better... a lot better. There is no consistency across the dashboard, no contrast in the colours used, and no indication as to what it's all about or what it is trying to answer. There must be a better way to show the same data in a clearer and "prettier" fashion. If we follow these top tips, we can do exactly that.

Tableau Dashboard Superheroes New York

Figure 1 - Superheroes in New York Original Dashboard

1. Before the Dashboard

Many of the best practices come even before the dashboard is produced, by improving each sheet to display the data in a more user-friendly way.

  • Choosing the best "Chart Type": The original "Offender" table used a line graph to show the number of offences committed by each villain, giving a false impression that the offenders are linked in any way; a stacked bar chart would display the unrelated data more efficiently.

 Bar Chart Tableau


The same stacked bar chart can be applied to make the "Occurrence Day" graph clearer, avoiding the hard-to-read area chart.

 area vs bar chart Tableau

Figure 3 - Before and After for the "Occurrence Day" graph

  • Removing clutter: The "Offender" chart has another problem: an excess of information. As the x-axis already shows the villains names, it would be best to dispose of the "Offenders" label by removing it from the Marks card as shown below.

Get Rid of Clutter Tableau

Figure 4 - Removing unnecessary labels

  • Using Maps: Maps are one of the best tools to use within Tableau, particularly since Tableau has inbuilt polygons to help users get the most out of their location data visualisations. Maps would be perfect for the "Zip" chart, currently displaying NY zip codes and their associated hero, showing all values simultaneously and taking advantage of the geographical nature of the data. Maps can also be improved by adding borders onto colours, customising the visuals to fit the dash's theme - but we'll get to that later.

Additional things to consider: 

  • Make sure axis are properly labelled, configured, and formatted, with no mismatched ranges.
  • Include units in the title.
  • Be consistent across sheets.

The dashboard so far:



2. The Dashboard

Now that we've sorted out our sheets, we can start arranging our dashboard. There are multiple things to consider when designing a dashboard's layout:

  • Use as few sheets as possible. There is no point in showing various charts with similar data. Currently, the "Offender" chart and the one below show identical data presented in different ways. The "Offender" chart is better as it follows our colour by hero trend across the dash, so we can simply remove the other one.
  • Structure charts according to what you want the user to see first. This is what I like to call the "added finesse". There are many different ways to arrange a dashboard, but my personal preference is to place charts in the order I want users to see them, from left to right and from top to bottom. I want the map to be interactive, and placing it on the top left corner makes it more likely to be the first thing the user sees, reads, and interacts with. I have also placed my filter/key at the top of the dashboard as these are key pieces of information, and then I have lined my sheets up by breaking the "Offences" into offender, occurrence day and occurrence hour.

Crime in Comic Books Dashboard Tableau


3. After the Dashboard

Now that we have a "presentable" dashboard it's time to move into the real personalisation stage, where we can add some "flair".

  • Colours: Of all the visual aids, colours are the most easily distinguishable and we should take advantage of that fact. The current dashboard's colour scheme is the "auto assigned palette", with shades very similar to one another and lacking in contrast. It's better to change that, either using your company colours or shades that go with the dashboard's theme.
  • Filters: Do you need your filters? Or can you make the dashboard more interactive? Tableau allows us to apply actions to every dashboard, which means that performing an "action" on one sheet will directly affect the other sheets within the dash, making the experience more interactive while removing clunky filters that might be cluttering our dashboard.
  • Style:  Much like with colours, we can change fonts and styles to suit our branding and message. Decide on your theme and find a suitable font type and colour to match, keeping the dashboard consistent.

With all the above things considered we have now a great looking dashboard as shown below.



Dashboard Checklist:

  • As few colours as possible
  • As few sheets as possible
  • Minimum space used by filters
  • Intuitive layout (left to right/top to bottom)
  • Consistent style throughout
  • Filters and actions

4. Extras

  • Custom Shapes: A picture speaks a thousand words. Why use a key with words when you can use a picture?
  • Custom Maps: The ultimate theme deserves the ultimate map. Services like MapBox work hand in hand with Tableau, allowing you to produce maps tailor made for your dashboard. You can check how to use MapBox in Tableau here.
  • Tooltips can provide additional information on top of what's already been visualised. There is a thin line between having useful and unnecessary information within the tooltip, so make sure it's relevant.

Below you can see the finished dashboard:

 To learn more tips and tricks on using Tableau, check out our blog.

Ryan Smith

About the author

Ryan Smith is a Graduate Analytics Consultant at Concentra. He studied Computer Networks at the University of Portsmouth and has a love of all things technical combined with a growing love of data. He believes that, together, these two can bring great insights into fascinating and exciting areas of everyday life. Sports is his main passion, and he enjoys visualizing all types of data from Football to MotoGP.

Related to Superhero Analytics: How to format your Tableau Dashboard in Four Steps