Welcome to the second part of my series on the ‘Visualisation Best Practices for Tableau or Power BI’. If you haven’t already had a look at Part 1 of this series, click on the link below.
Otherwise, we will be picking up where we left off – understanding pre-attentive attributes. Part 2 and 3 of this series will aim to answer the following questions:
- Part 2
- How could you use pre-attentive attributes effectively to create meaningful visualisations?
- How do you cater for audiences with accessibility needs?
- Part 3
- Which chart types should you use, and when?
How could you use pre-attentive attributes effectively to create meaningful visualisations?
There are ten main pre-attentive attributes that Tableau has identified, and these are also applicable to Power BI, see Figure 1 on the right.
Figure 1 taken from: https://help.tableau.com/current/blueprint/en-gb/bp_why_visual_analytics.htm
There is a sort of unspoken “hierarchy of importance” when choosing pre-attentive attributes. Basically, the first ones listed are the attributes that you should try to use before trying some of the attributes lower on the list. The list is as follows:
- Position and Orientation
- Length and Width
- Grouping and Enclosure
- Colour hue
- Colour intensity
- Size
- Shape
The reason behind this is due to the way our brains are able to process some of these attributes. Using too many attributes and using too many categories can increase the cognitive load as the brain tries to understand and interpret all the information being given.
For example, colour hue can be used to distinguish between different categories, but when you have too many, it can be overwhelming.
If you look at Figure 2 on the left, ‘Company Sectors’ are each given different colours. However, there’s a lot of information to process and this can become difficult to interpret.
Additionally, it can be hard to distinguish minor differences with Colour Intensity, while Grouping and Enclosure are also good for distinguishing patterns and categorising but there are limited use cases. Shapes are the least used as too many shapes can increase the cognitive load and create confusion.
Finally, size is also one of the more difficult attributes to judge unless there are significant differences. Let us look at some of the use cases for these attributes below.
- Position and Orientation
- Helps compare values, spot trends or patterns.
- Charts: Scatter plots, Slope charts, and Line charts.
- Length and Width
- Can compare magnitudes by easily spotting the tallest or widest bars.
- Charts: Bar charts, Bullet charts, Gantt charts, and Lollipop charts.
- Grouping and Enclosure
- Helps spot outliers or patterns.
- Charts: Heatmaps (highlight table) and Scatter plots.
- Colour hue
- Can differentiate categories and highlight specific points or values.
- Charts: Donut charts, Stacked Bar charts, Line charts, and more.
- Colour intensity
- Helps emphasise differences in numerical values.
- Charts: Heatmaps (highlight table) and Bar charts.
- Shape
- Can add additional information to charts using other attributes.
- Charts: Scatterplots.
- Size
- Helps with adding another dimension of information.
- Charts: Bubble charts and Tree maps (note: avoid these chart types if possible)
In summary, there are a myriad of pre-attentive attributes, and they can be utilised in conjunction with each other to highlight, emphasise or categorise the data. To achieve this effectively, you need to remember what each of the use cases are and ensure that you aim to not create clutter but clarity in your visualisation. We now need to consider accessibility when using colour as an attribute.
Let us dive deeper into this.
The importance of accessibility in data visualisations
When using colour, you need to take into consideration your audience. Your audience could include people with varying levels of colour blindness, and this would affect their understanding and interpretation of the visualisations that you create.
Take this map chart for example of colouring ‘States’ by ‘Profit’.
If you have normal vision, the differences might not be difficult to detect.
Check the next slide >
Now, look here at the same chart through a Green-blind/Deuteranopia lens.
How difficult was it to interpret this image?
Do you feel the same meaning could be conveyed?
What this means for your visualisations?
Well, you need to be very selective about when to use colour and how you use colour.
Be careful of your colour palette and understand your audience. Tableau has a colourblind palette that can be useful to use when you are not sure if your audience may need this adjustment. Otherwise, you can always try to use a website like Coblis — Color Blindness Simulator – Colblindor to check how your charts may be viewed through different lenses.
Keeping this in mind for the future will help you to cater for a wider audience.
Conclusion
In this blog, we have explored the numerous methods of ‘Visualisation Best Practices for Tableau or Power BI’, that assists you in ameliorating the method in which your data provides information to varying types of viewers.
After all, data is ubiquitous, but clear insights are unique; it is your job to create meaningful and authentic ways of interpreting data. This can be done by using coherent techniques that better visualise the chart types that are being employed, while applying strategic pre-attentive attributes, and catering for a broad audience.
In continuation, I will see you in the final part of this series where we will wrap up this series with a discussion on chart types in more detail and the specific ways we can use some of these chart types.
