Learn more about and join the Search Discovery Education Community!
In my 10+ years of working in the analytics industry, I’ve made some really terrible dashboards—the kind of dashboards that are more confusing than they are helpful and sometimes hardly survive for more than two weeks. However, through all these failures I found some common themes of mistakes that people make when building their dashboards. And, surprisingly, these mistakes are not about data – they’re about design.
A properly designed dashboard can represent any data and can clearly communicate a story in a way that makes someone want to use it. So if you’re a dashboard designer or a team leader helping to give better dashboard design feedback, listen up! You’re about to get a 301 in visual communication.
Why is visual design important in analytics dashboards?
It’s very important because we use analytics to help make business decisions, and if a dashboard is misinterpreted, bad business decisions can be a result. A good dashboard will:
- Effectively communicate information in a clear and concise manner
- Provide transparent business context
- Enable users to take action
- Encourage user engagement
Okay, let’s get into it and talk about a subject that many of us hold near-and-dear to our hearts—Wine! The US has nearly 10,000 wineries, with each producing at least a handful of wines, which leaves us 100,000+ wines to choose from. And what better way to help us choose a wine than to download all that data and build a Tableau dashboard! I’ll build a very basic dashboard using this wine data and then go through five steps of improving the dashboard with the tips I have learned over the years. Here’s the initial dashboard:
1. Progressive Disclosure
First tip! Consider progressive disclosure design principles when building your dashboard. Progressive Disclosure is the idea of only displaying the most important content in one view. Less important or supporting content can be moved to a second page. The idea is to not detract a user from what’s most important.
This dashboard is packed with data points, so I’m going to move three to page 2.
2. Choosing the Best Visualization
Choosing the best visualization is arguably the most important task of the analyst/dashboard creator as that visual will help present the type of data. Choosing the incorrect data can lead to misinterpretation and bad business decisions being made.
As a side note, choosing a horizontal bar graph works in many scenarios, however, if you only choose horizontal bar charts, your dashboard will lack visual interest. Keeping a reader’s attention is just as important as ensuring they are looking at the correct visual representation of the data.
Let’s consolidate some of these visualizations into one and use a map view to help provide visual context.
Hierarchy in visual design is the idea of arranging elements to show the order of importance. This traditionally means the most important content should appear in the top-left of each page/screen and elements that are larger on the page represent greater value to the reader. By laying out the elements logically and strategically, hierarchy can help to influence users’ perception.
I need to pull these numbers in the bottom-right of the page to the top for better identification and change the size of a couple of visualizations for better understanding.
4. Color Theory
Color is an extremely important aspect of dashboards. Color can be used to distinguish differences between metrics/dimensions or be used to support the design aesthetics. We all know that red is traditionally ‘bad’ and green is traditionally ‘good,’ but how do we decide which colors to use for my standard bar charts? How do I choose a complementary color?
The best resource for understanding color is to use a color wheel. A color wheel can teach you important design considerations such as which colors are complementary and which colors are analogous. This can help you start to build your color palette.
Our dashboard is looking much better, but there is something about the main color of blue that just doesn’t sing ‘wine’ to me. A change to a shade of purple would suit this dashboard much better.
Using icons and imagery in a dashboard can really make a user want to use your dashboards. Sometimes having a visual clue can help to create a better memory of the subject rather than a bunch of tables and graphs. Use imagery to create familiarity and to enhance the overall experience of using a dashboard.
To that point, I’ll add a logo to the top-right part of the dashboard for balance and set up dynamic image insertion to pull in images of the wine bottle I select. Now that creates a better user experience!
So there are a few tips on how to take your dashboard from drab to fab! Don’t be mistaken, there are PLENTY of other design considerations I could have taken to this dashboard (adding a legend, color-blind considerations, etc.) but I wanted to make this focused on the tips I was trying to share. There are a lot of other great design resources out there too, like the Tableau User Community, or one of my personal favorites is the dataisbeautiful subreddit.