Written by Matt Dent, BI Consultant at Data3
For context, this blog was written off the back of a conference talk at the annual Chartered Institute of Fundraising by myself and the fantastic Matthew Williams from Lloyds Banking group. The Chartered Institute of Fundraising is the professional membership body for UK fundraising.
I’m going to flip this question on its head and say; what are the consequences of not thinking about design?
If you are a BI developer, data analyst, insight manager, or just someone who wants to present data in a way; you’ll be well aware of the time it takes to get access to your data, and then to manipulate and transform it into the right shape and format. These steps can often take longer than expected which in turn means we have less time for visualisation and design. Not having time for design often leads to a rushed visual display, high cognitive load (hard to understand), and a disengaged audience.
So back to the original question. Better design = better understanding = more engagement = more usage = better business decisions
In this blog I’m going to speak about some design basics and principles that you can apply to your dashboards to help with your dashboard design. I’m a BI Developer, I’m not a design expert nor am I a user experience or user interface designer. I use some basic principles and techniques that can simply, easily, quickly and effectively transform your dashboards
First, let’s speak about web design (we can learn and apply a lot of this into dashboard design). A good example of this is to look at the evolution of the BBC news website:
A lot of dashboards I see look a bit like the 1997 dashboard. But you want to aim for 2020. Look at all the words in 1997, all the types of impact: underlines, bolds, changes of colour. Now look at 2020: containerised layouts, simple graphics, simple texts and titles, a nice colour palette. I’m going to be exploring these techniques and more in the following sections.
To kick off our talk we present a rather average dashboard. This was created in Tableau, and is the sort of dashboard you could create in a matter of minutes or hours, not days or weeks. It hasn’t been formatted or changed; everything is presented with default settings.
Think about the right chart for the right data, and the right interpretation of the data. For this we can use some guides. A great visual vocabulary is this one from the Financial times; it categorises data into different types and has some chart examples of how to display this data. If you are using a specific BI tool, such as PowerBi or Tableau, there are some great specific tool guides out there online.
- Use the same alignment across your dashboard. A good rule of thumb is to always left align.
- Think about the colours you’re using for your titles and fonts. Don’t use black! #000000, it’s harsh and technically not a colour. If you want to go dark use a nice grey like #888888
- For impact, just use one type. You can use colour, size, italic, bold, underline; just pick one and use it throughout, and definitely don’t use them all at the same time! I like to use font sizing to scale the fonts to the right size for impact. For example, a title needs some impact, just make it a bit bigger. A guide tool for this is: www.type-scale.com
- Font almost deserve their own blog. Go simple, use your companies brand guidelines, and use a font that will display on everyone’s computers, not something you’ve downloaded from online
- Remove clutter. Question if you need that chart title, those axis labels, those elements that require you to turn your head 90 degrees to read.
- Think about the ordering of your charts, order on the insight you’re trying to achieve, not just the alphabetical order of your categorical variables
- Think about the data you’re showing. Do you need all the data available since 1974, or would the last two years tell a better story?
Here’s an example of how we can put this into practice:
- Chart type – is a pie chart right for the data? Can you compare the sizes of the segments easily? Can you read all the labels? – In this case no, a bar chart is much more suitable.
- Do you need the axis? Is the title, right? What about the chart titles? How is the formatting? – In this case we can really simplify the chart, we don’t need an axis because we’ve got data labels, we don’t need an axis title because we’ve got Income in the title.
- We can also add some additional context – In this case we added the percentage difference to the previous year, to see both a trend and a value in the same chart
Here are some more examples of enhancing your visuals based of these principles:
Back to some more tips, tricks, and techniques:
Negative space is the space that is empty. The space between your dashboard elements. You should optimise this space, make it work as much as the space that is occupied.
- Empty space isn’t a waste of space, it allows your visuals to breathe
- Reduces cognitive load – not battling for your attention
- Allows us to remove borders
- Reduces space for cluttered dashboards and forces the builder to choose what’s the most important
A good example of this is used by Apple. Look how much negative space there is on their home page. It makes it feel professional, slick, and doesn’t distract you from what is important.
Here’s our negative space example, just getting rid of a few borders really allows the charts to breathe and be a part of the white background:
- For non-designers, we think this is the hardest element
- Our cheat methods include:
- Create a palette across you can use for all files; you only do this once (preference/themes file)
- Use your company’s brand guidelines (likely made by actual designers)
- We use https://learnui.design/tools/data-color-picker.html as a colour tool
- Colour is powerful, only use it where it means something….
Types of colour palette:
Use the right palette for the right data type. Is your data continuous, or is it categorical?
So, what happens when we apply all these techniques to our original default dashboard?
We have a nice clean, simple, yet highly effective dashboard. It uses consistency, uses impacted where needed, and adds context for reference and additional meaning. Lots of negative spaces keep the cognitive load down.
By using these 5 techniques, we can simply, easily, quickly, and effectively transform any dashboard design.
Better design = better understanding = more engagement = more usage = better business decisions
If you would like to watch our video on this topic, please follow this link: