Jeffrey Heer(University of Washington) has outlined 5 steps for wrangling data before you publish it:
Discover content and patterns.
Structure and format needed attributes.
Clean to eliminate meaningless outliers.
Enrich with data that adds context and meaning.
Validate by testing against logical constraints.
Brian Card and Mike Barry describe how they parsed a large MBTA dataset and iterated by visualizing frequently.
Intent is the “nugget of truth” that a visualization must make obvious. This visualization may be the only thing your audience knows about this topic. What do you want that to be, and what do you want them to do?
Your Audience should be comfortable with your tone and level of technical language. Make comparisons, allusions, and references that send the message that you get these people.
Cole Nussbaumer-Knaflic demonstrates how powerful your intent can be as you hone a visualization.
Randall Monroe combines a complex line drawing with captions that use the hundred most-common English words.
Analysis forms a feedback loop with Data Wrangling and Intent – If you realize that your data doesn’t tell the story you wanted, do you clean, manipulate, or add data, or do you re-evaluate what argument to make?
One way to make this process easier is to visualize early and often. You’re visualizing in the first place because it illuminates patterns and adds clarity. Take advantage!
Is your data trustworthy? Do you have enough data points? Do you need to show uncertainty? Are your outliers meaningful, or artifacts?
Nicolas Kayser-Bril discusses interpreting data. The Data Visualization Handbook offers additional chapters of advice.
An academic paper from the University of Leeds discusses illustrating confidence. The pictures alone are worth a look.
A visualization strategy has many parts, including:
What chart type(s) will you use?
What medium will you use (online, print, video, poster, etc.)?
How much variation will you show?
Every visual difference between your data points should impart meaningful and important information. You can show variation by changing position, color, shape, size, and existence (for animation or a series of images).
What chart type best conveys the message of your data? We offer advice for picking and building on the Charts Page.
Torsten Moller breaks down not only how data can vary visually, but which strategies should be used together, based on science.
Any designer has a toolbox that represents their own expertise, the expertise of others that can assist them, and the technical resources available in their organization.
You can maximize this toolbox by picking the right tool and getting the most out of it. Click “Tools” for tips on many of the most common software packages. Click “Implementation” for advice on how to develop, post, and host web-based visualizations if you aren’t an expert web developer.