The Dos and Don’ts of Dashboard Design
Quick Tips to Create an Impactful Dashboard
Many organizations and products use dashboards today. Dashboards are a great way to share data in a visual format. Tableau, Cognos Analytics, and Python are a few tools used to create dashboards. With all the different tools available, what does it take to make a great dashboard?
When a dashboard isn’t good, we as users quickly notice, which in turn often leads to low usage rates. When a dashboard is great and easy-to-use, we don’t take the time to think about it — we’re too busy finding interesting insights!
Over time, I’ve found a few tried and true techniques to making impactful dashboards. Below you’ll find a few DO’s and DON’Ts of dashboard design to help you with your next project!
DO: Add Text
Dashboards can contain more than visualizations! Adding text allows you to provide additional context to users. Depending on the tool used, the way you implement this will vary. Many tools, such as Cognos Analytics and Tableau, have text boxes you can add to a dashboard. I use text boxes underneath or near the dashboard title to provide users with instructions and a general overview.
Text boxes can be placed elsewhere on the dashboard to provide commentary or insights for the visualizations displayed. A few bullet points highlighting sample insights from the dashboard can help users start using the tool.
When it comes to adding text to dashboards, remember the KISS principle — keep it short and simple. Long paragraphs of text can deter users from utilizing the dashboard.
DON’T: Leave Default Font Settings
Many Business-Intelligence (BI) tools, including Tableau and Cognos Analytics, have default font settings. Personally, I have found that I rarely use these settings. The default settings aren’t terrible, but a slight adjustment can enhance the user experience! Adjusting the font size, color, or alignment can relay meaning to your users in a subtle way. A certain size or style can indicate a specific purpose to your users.
Below are a few ways you can modify the font on a dashboard:
- Incorporate a Tiered System for Font-Size — Consistent font-size helps users know where to look for certain types of information. When designing dashboards, I use the largest font-size for the dashboard title, the second-largest for chart titles, and the smallest for any additional text added to the dashboard.
- Bold Text — Bold all titles (including the dashboard title and individual chart titles), and any key words or phrases in textboxes.
- Italicize Text — Italicize text that users should take special note of as they navigate the dashboard — this could be special instructions for a certain group of users, or a disclaimer regarding the dashboard they are viewing!
- Add Color — Color is a great way to highlight key words, phrases, or information in textboxes that you want users to quickly spot. When adding key insights on a dashboard, I use color to highlight specific words or numbers related to the data, as seen in the example below.
DO: Add Filters
One of the key benefits of dashboards is the ability to slice and dice your data! The best way to do that is with filters! Many BI tools allow you to add filters directly on your dashboard, or filter your view using the visualizations.
When adding filters directly on a dashboard, I keep them in one place, so it’s easy for the user to find. I select locations where it supplements the dashboard, but doesn’t take away from the overall story. Placing filters at the top, right or left side of the dashboard, allows users to easily access the filters, while making sure the visualizations have the prime focus.
My personal favorite is using the visualizations as filters. I’ve found it creates an enjoyable, interactive experience for the user. They are able to create and navigate their own data journey, and dig deep into the dataset. For example, if you have a pie chart, a user can select a slice and all the visualizations on the dashboard will update to reflect that subset of data. Many tools allow for this feature, for example, in Tableau you can enable a Filter Action for each visualization, and in Cognos Analytics this is the default capability.
Note: If you don’t want filters to take up space on the dashboard, or you don’t want to use the visualizations as filters, many tools have alternative methods. In Tableau the Show/Hide button, and in Cognos Analytics the All Tabs & Current Tab filters are good alternatives.
DON’T: Randomly Place Visualizations on the Page
Where you place your visuals is important — think of what you want the user to view and interact with first, and place those items closer to the top. Consider alignment and visualization size — all visualizations don’t have to be the same size, which ones do you want to take up more space? What story are you trying to tell? Is there a way to place the visuals so that there is a natural flow?
It’s like reading a book, you wouldn’t read it upside down or out of order, so why expect users to do that for a dashboard?
When I’m trying to determine placement of visualizations, I create a few different options. First I start with a sketch, then create a few options with the tool I’m using to develop the dashboard. I also test how I would narrate a story to an audience. This allows me to see where the story naturally flows, and where it may need adjustments.
If you have several visualizations, but don’t know where to place them, consider adding tabs. These tabs can focus on specific visualizations or parts of the dataset.
DO: Include KPIs
KPIs (Key Performance Indicators), or key metrics, are some sort of measure that you can quantify. Total Number of Trades, Total Number of Claims, and Average User Response Rate are a few examples of KPIs. The metrics used should be of interest to the end user, since they can help with making key business decisions.
If your user only has 5 seconds to look at your dashboard, these numbers should stand out. Placement is key when it comes to KPIs on a dashboard. Place KPIs near the top for maximum visibility. Some locations I’ve found to be effective are on the top right corner, or directly below the dashboard title.
DON’T: Use Only One Visualization Type
A dashboard with only bar charts, or worse only pie charts, won’t leave as big of an impact. There are several types of visualizations that are beneficial for different data types. Consider the best way to display your data. While some repetition of visualization types is fine, try to incorporate some variation. Using only one visualization type on a dashboard can make it hard for things to stand out, and can lead to users glossing over key insights.
Common visualization types include horizontal and vertical bar charts, pie charts, word clouds, line graphs, heat maps, Sankey diagrams, and bubble graphs. For more on different visualization types, check out The Data Visualization Catalogue by Severino Ribecca.
DO: Use a Consistent Color Scheme
Dashboards with a consistent color scheme result in a cohesive look that helps build trust and credibility with your end users, while also setting the overall tone of the dashboard. Selecting the same colors for specific values or dimensions is a great way to increase cohesion and allows users to connect data points across a dashboard. Many tools also allow for custom color schemes — a great way for organizations to incorporate company colors into their dashboards!
If you’re not sure where to start, Data Color Picker is a helpful tool to determine the right colors for your data visualizations and dashboard — especially if you are using a tool like Python or R. This tool helps you test different colors palettes, single hue scales, and divergent scales.
DON’T: Feel Pressured to Stick With the First Design
Dashboard development is an iterative process, so don’t feel pressured to stick with the first version! The first version is rarely the final, and remaining rigid can reduce the overall quality in the long run.
Collaborate with your end users throughout the process and have them test out the dashboard as it’s being built. Observe their interactions, see what works and what doesn’t, and make improvements from there!
Follow these Do’s and Don’ts to elevate the design of your next dashboard!
Payal is a Data Scientist at IBM. In her spare time, she enjoys reading , traveling, and writing on Medium. If you enjoy her work, follow or subscribe to her list, and never miss a story!
The above article is personal and does not necessarily represent IBM’s positions, strategies, or opinions.
