In one the previous post I briefly outlined some of the things to avoid in dashboarding (data visualization). In today’s article, I would like to flip that post on the other side and explore what really is the recipe for a good dashboard?
This is not the ultimate guide – but, it’s something if well implemented will take your dashboards from good to great – and improve user engagement with your dashboards.
Know your audience
The best dashboard works for the right audience. Therefore, its necessary to begin with your audience in mind – is it a very busy manager with only 10 seconds to spare for your dashboard? Or is it for your supervisor interested in the nitty-gritty details about business performance?
Try as much to understand, who exactly will be viewing your dashboard – the most important metrics of interest to them and build your dashboard a round that.
Understanding your audience’s level of understanding – or presence of color blindness problems can help you address the same while designing your viz.
Consider your display size
Dashboard can either be designed for desktop, phone or tablet environments – or an automatic dashboard that resizes based on the device user is viewing on. Though, one may be tempted to go for the automatic view – sometimes, it can sacrifice the insights you’re communicating when resizing. That’s why its important to research the kind of environment your user will be viewing the dashboard on. And test your dashboard in the same environment, so to ensure the users experience is not compromised.
Limit number of views and colors
Generally, it’s tempting to want to cram your dashboards with many views and variety of colors in the name of designing dashboards. In most cases, this will only bring confusion in the minds of users – rendering your dashboard useless for not serving its purpose i.e. communicating data insights.
Always use limited number of views in your dashboards - 3 to 4 views will be good. And, limit the number of colors used to display your information.
Note; this include background and labelling color.
Eliminate clutter in your dashboards
Clutter can be anything that does not add value to your dashboard. It can be that title, legend or axis which is never helpful in interpreting the dashboard but only makes it appear busy. Try as much as possible to minimize if not eliminate completely anything which doesn’t help the user in interpreting the dashboard.
Make use of filters & tooltips to drive more engagement
One of the advantages of dashboards, and I mean interactive dashboards is the ability to enable users’ slice and dice data -meaning user’s do not just interact with static views but rather can change the level of display e.g. from course to fine level. And even leverage on hover for more details capabilities to interact better with data.
Whenever possible, try to make use of these capabilities to enable your dashboard user’s query data for that lower or course level question.
Test to ensure your dashboard load time is fast
Nothing will make your user’s get bored than a dashboard that takes minutes to load. With a society suffering from short attention span. No body will have time to wait for your heavy queries to get executed. Always plan your dashboards and minimize use of many/heavy calculations – most important, before publishing or sharing your dashboard with your users, test to see where queries are taking long to run – and find a solution around that.
The list can be endless, and different experts will have different opinions regarding this. However, practice will always prove more important than the theories. And the best way to test what works is by first implementing the ideas – get feedback from users’ and iterate until you get the optimal way of presenting data.
Thanks for reading.