Designing for the Data

Data visualization is more than just rendering charts; it’s about relaying your intention with what the data is trying to say.


Designing for data that’s expressed as web-based dashboards & mobile analytics apps often feels like a unique design genre, resulting from a potent confluence of graphic design, information architecture, and interaction design. Indeed, we’ve read canonical texts by Edward Tufte & others, issuing stern edicts that define proper, clear, essential information design practices– avoid “chart junk”, and so forth. Yet, there is so much more to it, that can unleash the significance of data in ways that enhance the creation of such forms. To get right into the heart of designing for dashboards and similar forms of interactive data visualization, I’ve gained certain insights and practices that can empower the designer, as described below.


* Pursue essentially “good data”: By this I mean not in terms of “data cleansing” or processing the data for specific formats, but in terms of the appropriateness of showing the right data, at the right time, in the right format, in the right medium/channel, to the right person. These necessarily interconnected levels of “right-ness” ensure that it is truly “good” data presented effectively for some human-based judgment or decision — given a context or situation, of course!

* Design with real, authentic data: Lorem ipsum simply doesn’t cut it, especially for complex information systems defined by IT admins or similar professionals. You’re likely dealing with strings of data not formatted “nicely” with extremely long, inhumanly readable names, or extreme numerical values, and non-proportional scales of magnitude that varies per object type or conditional state. Also, this raises issue of the integrity & accuracy of data for feedback by simply realizing IT Admins are too pressured or suspicious of “fake mock-ups” with suggestive, unrealistic content. Gotta keep it real to solicit valuable feedback and trusting support.

* Know what the data means: You’re not simply a Kinko’s style “formatter of data”, churning it out by the order, but someone obligated to ensure an accurate presentation of the data’s value. So, you must truly understand what the data means, chewing it up and ripping it apart to interpret the fundamental purpose and utility. Is that data even needed or is something else more effective here? Why does this data exist at all, how did come to being, what is the purpose for the data’s “consumer”? Is there “a social life of information” that extends beyond the screen or device? This helps characterize the data’s magnitude of significance. And as the designer in pursuit of “good data” referenced above, it’s your deep responsibility to capture and translate that meaning for the data’s intended consumers.

* Question the “What Ifs”: Diligently ask about the assorted, yet complementary set of states, conditions, situations, use cases, that can impact the presentation of the data and how or what it needs to communicate. What are all the dependencies, and assumptions being held by the user and the system, and how do they interoperate? Are there certain elements of the product system that impact the data latency, quality, freshness, and precision? Chances are yes, so identify them, learn how they interact, what the timing and sequence of data impact is, and who on the team owns those elements… thus, they can help you make effective design decisions, whether it’s a status update or a error message. Yes, it all trickles down and it’s up to you as the designer to work it all out!

* Engage with the back-end data team: Working with the back-end folks is critical to ensuring the the most appropriate UI is being shown in the proper context and timing, due to sequences of operations and trigger points for failures, or any other dependencies. They know operations and behaviors that may impact what is shown or why show something at front-end. You’re the one to provide the context and cue for them to understand that, thus fostering a productive two-way dialogue that keeps everyone in the loop and continually honing the product, the data, and their overall quality of experience.

One clap, two clap, three clap, forty?

By clapping more or less, you can signal to us which stories really stand out.