Design and data may seem like two unrelated fields, and in some cases they are. The truth of the matter though, is that design is also a powerful tool that can transform how we understand a number of different fields –– and these include data science.
Now that Big Data is growing in reach, industries are setting a high demand for data visualizations that are built on design and interactive elements. Good old visualization tools, like line graphs or pie charts, are no longer sufficient for complex datasets, especially given how important data visualization layouts are to the interpretation and analysis of information.
As the volume and scope of data continue to grow, design will only play a larger role in data presentation and analysis. To illustrate, here’s what you need to know about the importance of design in data visualization.
In several of his books, as well as in training and presentations, Edward Tufte has provided insight into the very creative and diverse ways that people have visualized data over the years. Of course for most of human history, data was presented in “static” form: print, as opposed to web or interactive formats, and that has been Tufte’s primary focus. Yet the core visual elements of print and web content, i.e. the mapping between data and design elements, have much in common.
In “The Visual Display of Quantitative Information,” Tufte cites an incredible image by Charles Joseph Minard, which presents data about Napoleon’s foray into Russia in a very informative way.
As Joanne Cheng has described, “Minard’s graphic is quite clever because of its ability to combine all of the dimensions: loss of life at a time and location; temperature; geography; historical context; all into one single graphic.” Reading Tufte’s book, you can see that data visualization has a long and distinguished history: a single 2-D image can provide deep insight into complex information.
As computers first came on the scene, data visualization was initially quite crude in comparison to Tufte’s greatest examples. Yet computer-generated data visualization offered two fantastic new capabilities:
With recent advances in web and database technology, things are coming full circle. It is now possible to visualize data with a high level of design, along with a subtle mapping between source data and its presentation, in a fully automated way.
When computers reached widespread viability, one breakthrough moment was the automated generation of data-generated graphics. Charting tools, such as Microsoft Excel and Adobe Illustrator, have become ubiquitous.
With the advent of the web, as well as the evolution of web standards such as JavaScript and SVG, interactive visualization has become more and more powerful. Frameworks such as d3.js provide the capability of drilling into data by clicking on regions of the image: this allows representation of data sets of nearly any arbitrary size.
Whether data visualization is automated or interactive, targeting print or web, design is a critical component that can make or break the power of the visualization.
To the untrained eye, fonts and colors may seem more like stylistic choices. However, designers know that these details can affect the audience’s perception of published content.
The same is true with regard to visualizing data content. Case in point: design director Howard Coale points out that serif fonts have more “visual noise,” which means that audiences take a longer time to understand numbers presented in this style, as compared to those in sans serif fonts. Coale also states that colors represent unique values in data visualization, in that each color family has a role in audiences’ understanding of the data values.
To create reliable insights, experts frequently need to make sense of thousands or even millions of data points. While conventional visualization tools can present the general patterns in these data, going the extra mile in design can highlight the correlations and stories hidden within the information. By visually organizing the data, the design is able to provide greater clarity about what the information truly represents.
When audiences and industry experts view published data visualizations, they don’t simply see statistics and numbers. Instead, they are able to understand how each data set affects the other, and what it all means in the big picture.
Design facilitates the interpretation and understanding of data. This turns visualization design into a powerful tool for data science professionals who leverage information as part of the process of recommending important actions. And since data analysts assist companies in identifying the latest trends and in making vital decisions, the added efficiency provided by quality visualization is crucial.
An organized data visualization layout improves a dataset’s appearance, but most importantly, it guides individuals to findings and conclusions more quickly. This leads to necessary adjustments being made when they’re most useful to an organization.
As you already know, design is more complex than it looks. If you’re publishing the content of your data visualization, it’s crucial to create a design that’s understandable and effective for the intended audiences.
However, you can also streamline this process and improve your organization’s data visualization through creative automation. To illustrate, Silicon Publishing was able to organize relational data about Royal Caribbean Cruise Lines passengers by leveraging the creative automation process of Adobe InDesign. Automation made it easier to personalize the layout, style, and text of the data within the document.
Aside from automating the design process, it is also important to choose the right visualization types for the datasets at hand. For example: in geospatial content, cartograms often use map distortions to present populations, while choropleths use map shades and patterns to show statistical values. With the right visualization design, it’s easier to accurately present the data to your target audience.
When it comes to data visualization content, design is more than an aesthetic choice. And now that there are more (and larger) datasets to be interpreted across industries, design has become a crucial instrument in the analytical process.