Effective visualizations are a combination of storytelling, message delivery, and communication. Some visualizations are at the level of art (and are even commissioned by companies), while others are more deliberate, carefully telling a story to the user through the orchestration of images, information, and/or animation.
With the exponential growth in data, today’s complex business problems require visualizations that go beyond pie charts and bar charts – those Ye Old data visualization capabilities of the past. To truly derive value from visualizations they must be dynamic, interactive, and actionable. They need to augment the message of the data without overshadowing it. And, as recent results from researchers at Harvard University and MIT have found, they need to be memorable, blending accurate data with influential features like color, density, and content themes.
Yet while it’s easy to get swept away in the “prettiness” of visualizations, it’s important not to forget their purpose. Visualizations should foster the easy consumption of large amounts of data: they should leave the consumer with an understanding of information, or give them new inspiration. The goal of a visualization – any visualization – is to provide a data consumer with the insights they need to make effective decisions quickly through a well-designed, visual end user experience. And, visualizations need to not only display information, but they must also be actionable – they’ve got to initiate a response that adds value to the business.
Consider this example: Aaron Koblin – the current Creative Director of the Data Arts Team at Google – monitored US flight patterns (141k of them, actually) over a 24-hour period, and then used the open source language Processing to crunch and munch the data and Adobe After Effects products to create an advanced data visualization. The Flight Patterns visualizations led to the Celestial Mechanics project – a “planetarium-based artwork installation that visualizes the statistics, data, and protocols of manmade aerial technologies” – by Scott Hessels and Gabriel Dunne of UCLA.
This award-winning project – while incredible (and super, super cool) – lacked the actionable piece that makes visualizations useful. Compare that to FlightAware, a global aviation software and data services company that tracks private and commercial aircraft in the US and abroad. As the largest flight tracking website in the world, FlightAware provides online flight planning, airport information, and airport fuel prices to data consumers. One of their tools, Misery Map, is a “real-time weather and flight data visualization that overlays Nexrad radar imagery on a map of the country, with red-green graphs showing the pain at major airports.”
Three Key Points to a Useful Data Visualization
To get the most value out of your data visualization – to ensure it’s not only communicating data but is doing it in a useful way – keep these three points in mind:
- Resist the urge to play with — or ticker with — what’s pretty in visualizations without knowing your message first. Just because something is pretty doesn’t mean it is effective – or useful. Be vigilant to cut out the “noise” in data visualizations, and focus on what’s important and how to communicate that effectively.
- Educate yourself in visualization, and know which visualizations are appropriate for communicating different data sets (such as time-series, comparisons, and multi-dimensionality). Think about the key relationships among data sets, and let this guide you to selecting the right visualization for the data you want to display.
- Understand storytelling with visualizations, and the importance of memorability. Data visualizations are about telling a story, using visual solutions to communicate a narrative that make the context tangible and digestible by the consumer.
As with any project, focusing on the end goal is important to guiding the creation of your visualization: a good plan equals a good design.
The 5 Faces of Visualization
In a 2008 blog post, FlowingData.com identified five different types of data visualization people and described them as such:
- Technicians are all about the mechanics. They have a strong programming background probably have worked with large databases. Technicians want to know how the visualization was made, and what makes it work.
- Analyzers aren’t too considered with aesthetics either and instead focus on the data. Analyzers want to know the relationships between variables, find trends, and discover complex patterns.
- Artists are concerned with what the visualization looks like – they’re the color and shape people. Artists are the people most likely to tell you that something is ugly.
- Outsiders are the ones with complex data sets, but aren’t sure what to do with it. They know the context of the data, but typically don’t know how to communicate it.
- Light bulbs are the idea people and to them it’s all about the big picture. They can lead projects and delegate work across a team. (Data Scientists often fall into this category.)
A sixth type of visualization person has been proposed: the hacker. This is the person who adds more and more features to the visualization, if for no other reason that simply because they can.
So, Which Are You?
When you look at these distinctions in the types of data visualization people, you can likely self-identify with one (or more) of these roles. Equally as important, you can also likely assign them to the people you work with, and the other members of your BI team.
Successful BI teams have a blend of these data visualization types – they’re a “best of” team that leverages the skillsets, personalities, and contributions of the technicians, analyzers, artists, outsides, and the light bulbs. By assessing the visualization people types in your BI team, you are better equipped to understand their strengths and know how to leverage their skillsets in projects, as well as where you have gaps in your team that need to be filled.
Further, understand the story, message, or usage before selecting visualization palettes. Make sure that you align the project with the visualization with the person for successful projects. Ultimately, remember that data visualization is a skill, and like other BI skills it requires education and discipline to avoid designing what simply looks pretty, and instead create a visual solution that tells a story and adds value.
About the author
Lindy Ryan is the Research Director for Radiant Advisor’s Data Discovery and Visualization practice and leads research and analyst activities in the confluence of data discovery, visualization, and data science from a business needs perspective. She also retains the role of Editor in Chief of RediscoveringBI Magazine. As Radiant Advisors’ Editor in Chief for three years, Lindy participated in in-depth discussions and analysis with industry thought leaders and vendors while maturing her position and perspectives in the BI industry.
Lindy has a B.S. in Business Administration and a M.A. in Organizational Leadership. She is currently a doctoral candidate, pursuing a PhD in Organizational Leadership and Strategy. Her dissertation research focuses on addressing the technical, ethical, and cultural impacts that have already and will continue to arise in a rapidly expanding big data culture.
Article originally published on InsideAnalysis.com. Repubished with permission from Radiant Advisorshttp://insideanalysis.com/2014/01/lets-get-visual/.