Why Designers Should Work With Real Data
Software applications don’t directly give you food, water, shelter, affection, or sex — the only thing they do is disseminate information that may lead to all these more basic human needs. Thus, graphic, UX, and software design is really about the selection and organization of information, through time. Information is everything, and you need to know what it is, what it means, and how to represent it in an honest and functional way to be a true asset to a company or society at large.
Designers aren’t scientists, mathematicians, economists, or even very rigorous researchers—and that’s fine. That in mind, most of them probably haven’t seen an Excel spreadsheet since their high school accounting class, and this is a big problem that’s plaguing the industry. In most of the shops or companies I’ve seen, and dozens of products I’ve worked or consulted-on, designers are either spoon-fed metrics they are supposed to visualize, or they simply assume what a particular bar graph or chart should read. As someone who has extensive experience both pulling, calculating, and visually representing data—I can tell you that the basic lack of statistical knowledge is detrimental to good product design, especially in the information age. This fear, ignorance, or oversight is especially apparent in the plethora of digital products that have functionless dashboards. Have a software product? Give it a dashboard!
Understanding the data you’re designing for will allow you to represent it in a way that’s functional for the end-user. It will force you (the designer) to become the analyst. Consider it a form of immersive ethnography. Rather than just choosing some high-level metric and mocking it up on a bar graph with pretty colors and fonts, force yourself (or your designers) to read the data first and determine what it is saying. What is important to your typical user? Is the sample size adequate? How does this representation give a particular persona the power to fix a relevant issue—to take the next step? How will this look with different sets and archetypes? All of these questions are important when choosing a specific visualization technique. Again, designers are responsible for the selection and organization of information—that’s really it. If you don’t know the many different ways of representing quantity, then your solutions will be more rudimentary.
How to Start Working with Data
With the growing use of prototype tools like Sketch, XD, and Figma — there is a definite trend for designers to move away from tools like Adobe Illustrator (and for good reason). All of these new-generation tools are fast and very simple to learn, but they generally lack the brute functionality of Illustrator as a powerful mathematical rendering tool. Especially the ability to create choropleth visualizations while layering and scaling data. When it comes to information design, Illustrator trumps everything.
Years ago I was working for an urban research foundation and was heavy into data visualization. This meant I was working with a lot of raw CSV files—importing data, analyzing it, and trying to figure out a way to make the data speak honestly and effectively. Subsequently, I had to learn Excel and ended up falling in love with the whole research process. I would get giddy when importing data and rendering charts to see what it all had to say. Learning how to pull, read, and visualize data is a unique skill that will propel anyone’s career as a functional designer. It also taught me the importance of research rigor and the scientific method, which has been tremendously valuable for me as someone working in software where egos can run high and social biases alive and well. In other words, proper research methods highlighted how crude UX ones were.
The beauty of Excel is that it works surprisingly well with Illustrator. You can create charts in Excel, change the fonts to AI compatible ones, paste them into Illustrator, then release the multitude of clipping masks. This leaves you with fully editable vectors that you can manipulate as you wish. This technique is also great for layering data to create rich visualizations that are perfect for analysts. Big data requires big visualizations and Illustrator gives you the canvas to do this. Always remember that ‘information designers’ (UX/UI/product/graphic) are responsible for representing the right data in the right way so people can make better decisions in their job or life. These are the people making multi-billion dollar decisions based on their interpretations, so there lies a heavy responsibility in the person organizing it. Moreover, working with data forces the designer to become the analyst—to become the user.
The experience of analysis is a process of juxtaposition—looking at different variables to understand relationships, patterns, and anomalies. The analyst is someone constantly moving their eyes. Design rendering tools such as Adobe Illustrator (using the blend and eyedropper tools) allows you to create stunning choropleth visualizations where one can ‘layer’ data, making it easier for analysts to gain critical insight on these relationships. Illustrator is also great for analyzing long trend lines because the user can set up their charts along with a significantly sized canvas, edit visual elements, easily annotate graphs, and send the graphs out to the ‘team’ for review. It’s also great for physically printing and mounting large visualizations onto a wall for analysis and annotation. Moreover, it is the task and responsibility of the designer to allow analysts to compare metrics very easily without making the canvas too complex. Visual complexity leads to cognitive chaos.
Real Data will Teach You Proper, Rigorous Research
The UX field at large is inflated by processes that are driven by hear-say, ego, and groupthink—not hard numbers and data. We frequently use qualitative feedback to justify design decisions when in fact, most times the sample size is too small for any insight to be considered ‘true’. The future of design will be about experimentation and evidence, not ego-driven qualitative research. Although the ego is necessary to formulate a hypothesis (or null-hypothesis if you’re doing it right), it isn’t healthy to assume or conform to an idea unless there is at least some evidence pointing towards its truth. Furthermore, exposing yourself (or your designers) to real data in real scenarios will inevitably result in better insight, strategy, process, and products. Dear designers, learn how to use Excel and get comfortable using it.
Design is the manipulation of stimuli and information, and as products become more complex, sound information design will separate successful products from poor ones. Not only does the layout, typography, color, and composition need to work, but the representation of the data must be fit for analysis. If you’ve ever made complex maps before, you can understand the difficulty in creating harmony among the different variables that a map is comprised of. It takes time, patience, and a willingness to investigate.
With the advancement of artificial intelligence and machine learning, one might think we may eventually not need analysts at all. Our computers and systems will make all of these decisions for us. This is a critical mistake that hardcore proponents of artificial intelligence make. Humans will and always want to be autonomous decision-makers. Even if and when AI starts to make hyper-accurate decisions about our purchasing behaviors, we will still need to inspect these suggestions, because free-will is one of our highest values and virtues regardless if you think it exists or not. At the end of the day, the human decides.
I’m Jeff Davidson
Give me your data and I will give you wings. I help companies visualize data and design meaningful digital products. Contact firstname.lastname@example.org for project inquiries. You can also get free design lessons on my site.