Seeing is knowing. More the visual appeal, more effective is the communique. The presentation of information becomes all the more important when you deal with business data as it needs to weave a visual story to keep the user glued. Data can be presented in several ways to make it look enticing and this blog by Karan Lahori shares the top ten steps to create intuitive and interactive dashboards so that it conveys the right message to the targeted audience.
Who wants to look at dull and drab data representations riddled with unimpressive tables and charts? No one! Dashboards are gaining stronger footage in representing data. As we move towards an information-driven market, where every organization, small, medium or large, is adopting data analytics to better understand the huge amount of raw data available and to say that the need for visual dashboards is imperative, would be an understatement.
With a wide range of tools and technologies available to satisfy this growing demand, here are 10 things you need to know for creating visual dashboards.
1. Think like the end user
The dashboard should be designed specifically for the type of user that will access and use the dashboard most frequently, and while it may seem like the obvious,, it goes a long way in implementing minor features, color combinations, levels of shading, font type and font size, etc. exactly the way the user likes (or similar to the reports/dashboards you are replacing, if applicable) for ease of acceptance. Who doesn’t appreciate intuitive presentations? If the dashboard is created keeping the end user in mind, adding functionalities and testing the dashboard will become simpler, which in turn will impress the end-users and reduce constant changes to the dashboard’s skeleton.
2. Find the right balance
While this is applicable only when the design is open to discussion, you need to find the balance between displaying a tab that looks clean and shows less (or high level, aggregated) data or a clustered tab that contains data at a more granular level.
3. Appealing design
While ‘Don’t judge a book by its cover’ is one of the most commonly heard idioms, it’s also true that the first impressions last the longest. A dashboard is developed to give a good understanding of the data available, at a glance. A dashboard is supposed to catch your attention and should not be just a dull representation of tables and text boxes, which is what a report is usually used for. Use all the aesthetic knowledge you have and make the dashboard as stunning as you can to give that great first impression!
4. Use popular widgets/chart types
There is a reason that you see the same chart types like bar, pie and gauge used in almost all dashboards because they are easy to understand and users can relate to it immediately. These chart types are fast and easy to view chunks of data which is exactly what you need in this scenario.
Another reason users want dashboards instead of reports is interactivity. Basic actions like filtering and drilling should be easy to perform. The users should also be able to get a good overview from the ‘Home’ page and then be able to follow links to other reports/dashboards or just be able to open interactive windows to see minor details and be able to return to the Home page based on their requirement. The easier it is to interact with the dashboard, the more enjoyable is the end-user’s experience.
While it is important to present a beautiful dashboard with great interactivity, it beats the purpose if it takes more than a few seconds to load the dashboard or switch to other tabs. Most often the end users are mid to high management personnel that value their own time and do not want to be kept waiting while the dashboards load.
7. Data formatting
Using the right data format wherever necessary is something that cannot be overlooked. Using the right currency symbol ($ vs £), date format based on the end users’ location (mm/dd vs dd/mm) or using percentage (or similar) values rounded off or with decimal points are small but important points that should be handled without the end users having to mention it.
8. Export options
While creating dashboards that look great on a web browser (or a mobile device) is the primary objective, it is equally important to ensure that the exported version looks clean and none of the sections appear broken. The ‘Export’ button should be placed strategically and the corresponding icons should be self-explanatory. Excel and PDF are commonly used formats and these exported files should be tested for formatting issues and interactivity, if applicable.
9. Get your spelling right
Grammatical/spelling mistakes are repulsive and should be avoided at all costs. Dashboards usually contain sensitive data, like sales/finance data of branches and their respective branch heads/managers, and no one likes to see typos or similar errors when dealing with such information.
10. Check data
Last but not the least important, no matter how impressive your dashboard is visually it is completely useless if the underlying data is incorrect. Make sure the data is tested thoroughly before it is deployed.
Hope these pointers prove helpful for your next dashboard. Do leave comments shall you find this interesting.
Here’s an example of a brilliantly designed dashboard:
Our fellow Saamaite, Hemang Bhavasar participated in Tableau’s global virtual hackathon – “Hacking open data: let’s get digital” #HackingOpenData”. From February 17th-23rd, participants from all over the world represented open data sets by using their visualization skills to demonstrate compelling stories around one of the five topics; agriculture, climate, education, energy and government.
Hemang chose Climate as the topic and utilized open data for “Water Source Risk in European Regions” from EEA (European Economic Area). His dashboard was one amongst the final 62 visualizations shortlisted by the Tableau Team and thousand others received as the nomination.
Kudos and Heartiest Congratulations Hemang !!
Access Hemang’s Viz here:
Result Link: https://public.tableau.com/en-us/s/blog/2017/03/wrap-hacking-open-data-lets-get-digital