ps:/ I’m a UIUX designer who helps in designing data visualization UI, making complex datasets more understandable and quicker to comprehend — I’m not a data analyst or statistician. I designed and passed the designed sheet to the developer to deploy it.
In 2023, my Business Unit (BU) took on me to improve the existing Tableau dashboard. Despite data visualization not being my specialty, I seized the opportunity to explore this new design area.
Like other designers, I start my process by understanding the requirements and looking for dashboard design examples. There are many examples you can find on Pinterest, Behance, or other design websites. However, the challenge arises when we’re checking the feasibility of the design for development.
Here are my tips and lessons learned:
Understand your audience and requirement
Before you begin designing the dashboard, it’s crucial to understand your audience. This knowledge aids in clustering and sectioning the data. For instance, if your audience consists of an Account Manager, you should know what they are interested in seeing. Similarly, if you’re designing for lab analysts, you need to understand the attributes they want to view. Knowing your audience ensures that the dashboard you create effectively achieves its objectives.
Once you receive the dataset, consult with your stakeholders about their dashboard preferences. Align their requirements with your research on common audience expectations. This will help you better understand their needs.
Clear up the requirement from the top view until narrow down the visual.
Understand your data
Next, once you know the expectation, learn about the data visual. To achieve the objective, understand the purpose of each chart available- What would you like to show?
Do you want to compare total revenue from 2020 and 2023 or do you want to see the trend of the revenue? Maybe, you want to see the distribution of rainfall in Mexico monthly? — Understand your data
Don’t immediately dive into intricate and elaborate graphs. Instead, begin with basic graphs such as bar graphs and pie charts. Once you understand their usage and match it with the requirements, you can then enhance your graph design. Remember, it’s best to start small.
In my experience, complex graphs can be hard to understand unless the user is proficient in interpreting data. However, typically only 1 out of 10 team members can understand these complex visuals. Therefore, it’s highly recommended to design a graph or dashboard that is accessible and understandable to all types of users.
My lesson learn: During the process, my Business Unit requested the use of numerous boxplots to visualize the quartiles (low/mid/high) of the survey data. However, upon publishing the dashboard, it became clear that the boxplots conveyed excessive information unnecessary for the user. Therefore, it’s important to understand your purpose for using data visualization.
Understand the technical limitation
While there are numerous analytics tools that may offer more advanced data analysis features, in my case, we focus on Tableau limitation.
Avoid using the Behance dashboard design as a reference. Instead, consider the app dashboard example for inspiration. For our purposes, we need a design suitable for a data set in the Tableau application. While the designs shared on Behance are indeed beautiful, with perfect 8-grid rules and sufficient white space, they are not feasible for implementation in Tableau.
We’ve decided to refer to the Tableau library for design examples. If you understand the requirements and know what type of graph you need, search for a similar example in the library. You might not find a 100% match, but at least you’ll have a reference to start with. If a developer claims there are limitations preventing the creation of the desired design, provide them with a URL example from the library. The Tableau library not only offers design references, but also provides the code to achieve that design. It’s a win-win situation
Understanding technical limitations can be a bonus as it helps to avoid repetitive design. For instance, I’ve noticed that it’s challenging to display data dynamically in Tableau.
When I began designing, we suggested a filtered data set that would allow users to show, hide, and add comparisons to the graph. We imagined that if the default data compared two items and a user wanted to add one more item, the display would adjust to three columns. We also considered the possibility of a dynamic column increase. However, we soon realized that this was a limitation in Tableau. Developers need to set the maximum number of columns from the start and can only display an additional column when the user wants to view it. So, we decided to work around this by limiting the number of comparisons and columns.
Another example:-
Originally, we planned to display the “Productivity in the Lab” area using this UI. However, we found that it wasn’t feasible to link this area to the list of Laboratories in the next section using this UI. This discrepancy deviates from our initial objective.
In conclusion, I learned those 3 main points during the four-month period of designing in Tableau. Understanding these has allowed me to speed up the design process and achieve objectives with fewer iterations. To recap:-
- Understand your audience and the requirements
- Understand your data
- Understand technical limitation
I hope this article may help others who are interested to help data into visual form. Goodluck!
