Describe Data Graphically
The alphabet was a founding technology of information. The telephone, the fax machine, the calculation, and ultimately, the computer are only the latest innovations devised for saving, manipulating, and communicating knowledge.”
- James Gleick, The Information, p11
We live in an over-saturated data-sphere. Try reading a government report some time — heavy on the facts but lean on the figures! For most people, synthesizing data through words or tables is actually quite difficult. With the explosion of data since the digital revolution, we’re increasingly faced with synthesizing large quantities information. As an overwhelmingly visual species, we humans turn to visual means to communicate complex data sets. Often called infographics, data representations may be generated from numerical data (graphs), spatial data (maps), or abstract / formal data (diagrams).
In this How To, Dan discusses the distinctions between types of data, how to represent them, what it all means, and tips to create your own infographics with the goal of telling a story. This post is supplementary to a lecture given by Dan, in the course Sustainable Material Assemblies at the Boston Architectural College on February 4, 2014.
Particular parts of the lecture are dispersed through this post, or you may listen to the entire lecture here:
STEP 1. WHAT TYPE OF DATA IS IT?
Infographics may be generated from one of a few types of data. Understanding what type of data you have will help to determine the proper form of representation.
Infographics by John Bernardo
Numerical Data: Graph
Ultimately, just about all data may be represented numerically. But here, we refer to data that explains proportional relationships between discrete entities that may be abstractly represented through a graph. Typically this data is found in spreadsheet format, and may often may be graph-able directly within Microsoft Excel.However, Excel (or google docs) have limited graphic capabilities. Once you have created a graph in Excel (not google docs!), copy and paste it into an Adobe Illustrator document for graphic attention. If your data set is more complex than can be represented through a simple graph, consider working within a programming language such as Python or Grasshopper for your representations.
Here are a few examples of simple graphs that seek to tell a story. What do they tell you?
Spatial Data: Map
If your data set includes locations in space, mapping is likely the representational tool of choice. Maps may be at any scale, and of any dimension (maps may be of atomic relationships, at the scale of the city, mega-region, or the galaxy), but fundamentally identify elements as they relate to each other in time-space. At the scale of human habitation on earth, we turn to GIS (graphical information systems) as the primary method of compiling digital spatial data. When combined with a series of other applications such as Google Earth, the Adobe Creative Suite, and Rhinoceros, we can create robust and richly detailed maps. Look out for our next SC How To that drills into this workflow!
Here are a few examples of maps:
Abstract or Formal Data: diagram
While graphs and maps are fairly easy to conceptualize, diagrams are more elusive. Diagrams do not merely present information graphically, they MAKE EXPLICIT the relationships between various elements, abstracting certain conditions, or making graphic traditionally non-graphic entities (time for example). Therefore, diagrams may also contain numeric and/or spatial data representations. Diagrams also represent formal data, such as geometrical or biological forms.
Here are a few examples of Diagrams:
STEP 2. WHAT DOES THE DATA MEAN?
Data representations seek to not merely illustrate information, but to create knowledge. To actively illustrate RELATIONSHIPS between pieces of information, data representations are more than nice looking graphics. They must tell a story. In this graph below, it’s difficult to discern what matters. In the second version, important months are highlighted with a text annotation to explain the information. The graphics synthesize the information, making it possible for the audience to formalize it into knowledge.
Avoid Chart Junk
Edward Tufte is a master of data representation, and Chart Junk is a major culprit. He was particularly critical of Microsoft Powerpoint, which offered graphics tools for folks like doctors and lawyers who are not always so graphically inclined.
Image Source: Edward Tufte
Presentations by the inexpert may include a barrage of graphic styles for different pieces of information, making presentations quite difficult to understand and obscuring the meaning of the data. Therefore, the number one rule in creating infographics is:
Every graphic element in the drawing should mean something. If it doesn’t convey some piece of information, then it’s probably unnecessary.*
- Recently the discussion over Apple’s use of skeumorphism addressed this issue directly, as many of the graphic elements in the pre-iOS 7 operating system did not serve any purpose for the action performed. Initially the graphics did in fact serve a very legitimate purpose, conveying the meaning of abstract digital graphics to consumers who had never touched a magic screen like the iphone before. However, now that people are comfortable with the new media, the meaning is lost.
STEP 3. SERIES AND PARALLEL REPRESENTATIONS
Sometimes a simple, clear set of graphics are required to communicate the information and be understood by the reader as knowledge. Other times, the representation may serve to synthesize a few disparate pieces of data into knowledge on the page. I like to compare these techniques to electric circuiting.
Show each data point as an individual graphic, as in the wind data for Boston, MA on the left in the image above. Although not absolute, series-based representations tend to work well when time is a variable, such as when recounting narrative, showing phasing, or change over time.
Superimpose data, bisecting multiple representational types. The wind chart on the right includes multiple data sources superimposed over a polar graph of a compass (which could be considered a data source as well). Parallel representations are useful when multiple layers all contribute to a complex reading. Maps are generally parallel representations.
STEP 4. ITERATE!
As with any design process, iteration will lead to a deeper understanding of data, and therefore increase the quality and legibility of the graphics.
The first critical in any design endeavor is to sketch the idea by hand. The relationship between hand and brain is strong, take full advantage of it to formulate at least a general direction before jumping into the digital world.
Then, go digital
The images below are a series of iterations of a timeline I designed for my thesis research on Haiti and digital climatic analysis and design processes. The first iteration on the left creates a diagonal expression that does little to support the content. The middle has a distracting set of grey. By the final iteration, the supporting events converge along the popularly understood narrative history of Haiti.
STEP 5. BE CONSISTENT
As you create a body of images that conspire to tell one coherent story, a consistent graphic language is the primary means by which to stitch the story together. Color choice, line weights, fonts and texture all contribute to create relationships across drawings. Set up a style sheet as quickly as possible, but inevitably your final drawings will be of higher quality than your first. If you plan to show early drawings in a final presentation, make sure to return to the early drawings and update their graphics to match the tone and quality of those produced since.
In this demo I take a snippet of data from the wikipedia article on brick, and create a simple graph, which when compiled in with a series of other data representations, could tell the story of the brick industry, and it’s relationship to a particular architectural project, for example.
You can download a copy of the Adobe Illustrator file (CS6) here.
Being a designer requires a willingness to lose a little control – to let go of assumptions and preconceptions – to just sit with ideas, information, content, or data. To stew in it; to bask in it’s impartiality. Through working with it, molding it, we begin to tell a story, to take impartial content and give it meaning, a subjectivity and humanity. This is the ultimate goal of infographics.
Good luck data mining, post questions, works-in-progress, or comments below!