How Does Data Visualization Work?
We work with data visualizations so much at Visually, that we often forget not everyone is as familiar with them as we are. One of our goals is to increase the world’s visual literacy. An important step in that is teaching people how visualizations work.
As kids, many of us learned about pie charts, bar charts, line charts, and maybe even scatterplots. We might have even used colorful candy to make some of them. But this doesn’t mean that we learned how or why data visualization works.
Data visualization is an amazing tool. The data we deal with daily would be almost entirely inaccessible when locked up in numerical formats. Luckily, data visualization can help us to extract information, insights, or even knowledge from that data. It relies on the remarkable human visual system that turns visible light into meaningful semantics that inform our decisions.
The Human Visual System
So how is information communicated to our brain? At the very beginning of the process, our eyes collect light reflected or emitted from the objects around us. Our corneas and lenses in the front of our eyes focus it onto our retinas at the back of our eyes, where rods and cones sensitive to different wavelengths of light measure brightness and color. The information produced by our eyes at this point is a raw image, similar to a digital image that any digital camera takes.
This raw image is further processed in our brain to build a data structure of objects with visible boundaries. At this point, these objects only have visual and spatial definitions. They don’t have names, properties, or any other information connected with them. The process of creating this object structure uses a set of rules known in psychology as Gestalt laws.
Once our brains have things grouped into objects, we can connect them with names, or other meaningful concepts and knowledge we have about the world. The whole system allows us to interpret the world around us at lightning-fast speeds in meaningful ways. Normally, we use it to learn about the concrete, tangible world around us, but with visualization, it can also be used to learn about the abstract.
Since our visual system evolved to help us survive in the tangible world, data visualization relies on metaphors to the physical world to communicate the concepts that exist in the data. Pie charts use a pie metaphor to show a part to whole relationship. Continuity is indicated by the connected line in line charts, while bar charts, with their individual bars are great for discrete categories. Arguably, every data visualization relies on a metaphor of some sort to help the viewer understand what they are seeing.
The reason data visualization is so powerful is it shows us views of the world that we wouldn’t ordinarily get. It is a lot like standing at the top of a mountain, looking down on the valley below (in fact, that’s exactly what maps do for us). The vantage point that data visualizations provide for us gives new context to frame things. Data visualizations help point out patterns and order that exist in the data.
One of the best examples of this is Anscombe’s Quartet. The numbers in the quartet are all statistically indistinguishable. They have the same mean, variance, correlation, and linear regression, but when they are visualized, they have definite patterns that we can see.
But using the right visualization is a big part of the process. Picking a visualization at random isn’t guaranteed to show you the insights you are looking for. Visual analytics tools like Tableau support a process of data exploration using visuals. For more custom visualizations, designers use sketching and iteration to develop the final format for presenting the information. This exploration stage is critical to finding the right metaphor for displaying the information you want to show, and for finding the information stored in the data.
Once the exploration phase is done, it is time to present the findings. This presentation phase is what Visually is, and what we do every day. Just like any presentation, beautiful, polished work is important for presenting data. It helps people focus on the insights they are being presented with, and makes sure the message is communicated clearly.
Drew Skau is Visualization Architect at Visual.ly and a PhD Computer Science Visualization student at UNCC with an undergraduate degree in Architecture. You can follow him on Twitter @SeeingStructure