As a UX Designer supporting a variety of clients, I am not restricted to working on one design type. I need to be flexible, since I may use established designs such as the Government Digital Service (GDS) and unique, client specific libraries. Whilst some designs are static, others will need to be more dynamic.
Data-driven design falls into the dynamic camp. It is a design style that adapts on the fly for the user when they load the screen, so that it can tell a story. Because of this, it requires consideration when designing.
Data-driven design – A different challenge
Data-driven design displays dynamic data rather than static information. This means that when the data changes, the design needs to respond without being re-designed. To shape the adaptability of the design, data-driven designs require a deeper understanding of the underlying dataset by the designer.
An example of a data-driven design is a table with items, where the items might grow to an indefinite number. If the table design can only cope with ten rows before it loses its usability, accessibility, and visual appearance, it’s not adapting to the possibility of the data changing.
I love data-driven design. It presents a different type of challenge. It is a lot of fun researching, designing, and evaluating a forever changing dataset to find the perfect solutions on how to best present it back to its users. Good data-driven design will give its user information and tell a story for them to use when building their understanding. It will also allow them to base complex decisions on easy-to-read history and trends.
Getting the best out of data-driven design
How can you get the best out of data-driven design? Well, in my experience, it is best to keep it simple. Here are three rules that I follow when I am designing for information that’s constantly changing.
- Research the dataset
Before settling on a specific design, look at the data your design is meant to feature. Is it limited or growing? If the data is constantly being added to, then your design needs to allow this to happen while still maintaining its usability. Understanding the dataset is key.
For example, if you are reviewing a table showing the number of incidents logged, grouped by type, then questions to ask include:
- Is the type of incidents fixed, or are people adding new types?
- Will the table eventually need to accommodate an indefinite number of types?
- Are there limitations of the real-estate available on screen?
- Will people interact with the user interface, or will it display as read only?
Depending on the answer, you might want to change how you design the table and incorporate tactics such as infinite scrolling, page options or limited static real estate.
- Research available diagrams
Data-driven designs often display data and statistics. When working with data and statistics, there are several different diagrams to visualise your dataset. They will all have different dimensions and work differently depending on the dataset, so it’s good to know your data when selecting what diagram you might want to use.
A simple bar graph works well with just a few or many bars. However, a line graph can easily look cluttered and decrease usability and accessibility if the number of items is too many. Diagrams such as bar graphs, line graphs and bubble charts can all contain multiple groupings, but a pie chart might work better when limiting the number of groupings needed.
- Test your design
It is always crucial to test your designs with users and it’s just as important to test them when working with a forever changing dataset. Testing your design with different numbers of items in your dataset is the key when working with data-driven design, both in your uncoded and coded prototypes. If you can, import real data into your coded prototype, as this will give you great insight into how your design will render. It is fascinating to see how the visual appearance of a design changes once your dataset grows.
If you have a question for Jenny or the Triad team, please get in touch.
