'The greatest value of a picture is when it forces us to notice what we never expected to see.' - John Tukey, American Mathematician
The collection and analysis of data is a wonderful thing, but if the insights these efforts yield aren't acted on, they are little more than a very expensive waste of time. In order to influence the behavioral change in your organization necessary to create a data-driven culture and ensure that action is taken, it is vital that patterns revealed in the data are either communicated in a convincing way, or that data is presented such that patterns can be easily found by layman users.
Data visualization is, however, easy to get wrong, and if done badly, it will have a completely adverse effect. Fortunately, the practice has evolved dramatically in recent years, with knowledge around the area advancing significantly as organizations have realized its importance and devoted resources to perfecting it. They are also able to utilize an increasingly impressive array of software, moving away from basic bars and charts and towards interactive spectacles that really engage and persuade the audience, whether it be decision makers in a boardroom or children in a science museum.
There are three key things to remember when creating any visualization. Keep it simple, tell a story, and engage the audience. Focus on providing only the information necessary in telling your story. Mico Yuk, founder of BIbrainz, argues that, "Great data visualizations tell stories. Not just pretty stories, but stories that provide insight, outcomes, and actions. Telling a good story is not complex, but the simplicity is where most fail. A great story does not tell you everything, just the things you need to know." Speaking at the Data Visualization Summit in London, Chi-Yi Kuan, Director of Business Analytics at LinkedIn, agreed, adding that "an effective visualization should combine both art and science. For art, not just having a creative mindset for your visualization, but also keeping your design in a style of neat and simple to show those patterns or relationships in the data. For science, it is not just to develop various types of charts, but also make complex data more understandable for the fact of the past and the prediction of the future."
Here are six visualizations that you should use as inspiration in your own data efforts.
Content meets form
Image via The Guardian
Data-based reporting is now firmly ingrained in the newsroom. According to a recent study by Google News Lab and PolicyViz, 42% of reporters use data to tell stories more than twice a week, while 51% of all news organizations in the US and Europe now have a dedicated data journalist - and this rises to 60% for digital-only platforms.
This is inspiring some of the most engaging and intriguing data visualizations, as journalists seek to appeal to a mass audience. The Guardian, in particular, has become expert in marrying content and form to drive home what their analysis of datasets, such as the Panama Papers, exposes. The above image showing the number of sexual harassment incidents that go unreported may be simple, but because the imagery used it is also incredibly powerful, it really drives the point home.
Image via NY Times
Interactive data visualizations are a great way of engaging the audience without the need for explanation. The above graphic details Tommy Caldwell and Kevin Jorgeson's 19-day free climb in Yosemite National Park in California, complementing the story and enabling the user to fully appreciate the scale of the achievement. These are easily created on websites like infogram.com. They also allow the business user to explore the data themselves and reveal their own insights.
Show the journey
Hans Rosling once said that, "Having the data is not enough, I have to show it in ways people both enjoy and understand." Rosling was the master of engaging his audience, both in terms of the innovative graphics themselves and his own unique presentation style. He didn't restrict himself to presenting data in a digital format either. He also used a variety of props - lego bricks, cardboard boxes, teacups - alongside his vibrant, animated data visualizations.
As demonstrated in this video showing poverty levels over time, Rosling was so good because he was able to hold the audience's attention and walked the audience through what the data was saying. He explains the horizontal and vertical axes. He then explains the starting point and gives context before going on to show you what has happened. Most people make the mistake of unveiling all of their data visualization at once — without offering any direction about what to look for or explaining the variables. They then move straight on to their own conclusion leaving the audience bewildered and lacking any understanding of how it was reached. This is a particularly pertinent lesson for business users looking to get decision makers on board with their insights. It is not enough just to wave a pretty graph at them and tell them what it says, you need to show the journey.
— Ed Hawkins (@ed_hawkins) May 9, 2016
Image via Ed Hawkins
GIFs aren't just for teenagers, they are also an engaging and snappy way of demonstrating how something changes within a given time frame. The above shows how temperatures spiraled between 1850 and 2016. It is impactful, and they are relatively easy to make.
The importance of color
Image via morgenpost
Color is a vital aspect of a good data visualization. Color is a powerful tool and should always be used to convey a message, it is not just there to make it look pretty. A heatmap – adding colors that vary in intensity to show relative performance – is a great way to help readers more quickly process the imagery and uncover patterns. The above image is The Berliner Morgenpost’s EuropaKarte, a detailed map which viewers can explore to uncover detailed insights into the population growth and decline in Europe. The fields vary in colors and their intensity, with vibrant orange symbolizing the biggest growth and dark blue the biggest decline.
Immersive new technologies
Immersive data visualization offers easier pattern recognition in big data sets and more intuitive data understanding. Unlike conventional data visualization models that rely on two axes in a spreadsheet, VR allows users to walk around and look at the data from multiple angles, comparing any number of different factors at the same time.
VR works because it focuses your entire field of vision, allowing you to concentrate exclusively on the objective. In early experiments, researchers found that users who interacted with the data through VR reported better retention of perceived relationships within the data than when viewing it through two-dimensional data visualization tools. According to SAS software architect Michael D Thomas, we cannot process more than 1 kilobit of information per second when reading text from a screen. By being ‘present’ in the data, you can get a true sense of scale which is impossible to achieve when viewing the data on a desktop screen.