What Data Visualization Means
Visual story-telling with data combines the best of art, science and engineeringThe Art of Data
There are many ways to define what data visualization is, but like Justice Potter Stewart’s famous opinion on obscenity, you’ll know one when you see it! Indeed, visualizations often rise to a form of art. For instance, consider the images shown here; they are beautiful and have esthetic value in themselves, without necessarily conveying any special insight.
To us, visualization is an art too: the art of showing data in another form so that insights or meaning are easier to understand. Visualizations are often graphical or image-based, but they could also be text or numerical. The transformation of the data into visualizations often sheds redundancy or complexity to reveal an otherwise obscured meaning.
Technology in the Service of Data
We feel at its best, good visualization is also good story-telling: a visualization engineer has a point-of-view to communicate. He or she builds that story from data and a set of artistic and technical tools and, like a writer, constructs a narrative to lead the viewer to a certain view of the world or perhaps even a conclusion. He may in fact elide or suppress irrelevant information. But a visualization engineer cannot make things out of whole cloth; the data must support the point-of-view or narrative. In this respect, creating visualizations is more about letting the data tell the story in the best way rather than telling a story by picking the data.
Nevertheless, we think it’s essential to master all the technical tools and details that are a part of developing visualizations; yet, that’s only the first step. We go further: to draw out the insights so that others can readily see them too.
Visualization: Handmaiden to Data Science
To successfully draw out the relevant by crafting the optimal visualization, an empathy for the subject at hand is ideal, if not required. Without that, the story-telling and insights may well fall short of what would be most interesting to the intended audience.
Today, visualization is no longer a luxury; it has become necessary because otherwise large and unwieldy data sets become incomprehensible to most people, the epitome of information overload. It turns out that evolution has bestowed upon us humans a very highly developed image processing and visual pattern recognition capacity. The art of visualization attempts to transform data so that this powerful human capability can be easily harnessed. In this way, visualizations can improve humans’ decision making by using our perceptual system, and in particular the ability to quickly identify visual patterns.
Guidelines for Compelling Visualizations
Though not preconditions, there are guidelines that have helped us design successful visualizations:
- Present data in a compact, consistent and coherent form.
- Show information from various viewpoints.
- Provide information at various levels of detail, allowing slicing the data or drilling down into it in different ways.
- Support comparisons over time, space and other dimensions.
- Enable data-based storytelling.
Underpinning these considerations is an abiding empathy or respect for knowledge about the domain and its stakeholders. Without that, the story-telling and insights may well fall short. Looi Consulting takes pains to understand clients’ problems and knowledge domain; then, we bring to bear the combination of engineering, data science and great story-telling to build our visualizations.
LCL helped us develop tools to better present the data we had been collecting for years. In just a few months, they designed and delivered a solution that yielded new insights and allowed us to view the data in ways we couldn’t before. The visualizations they designed have led to new ideas both in presenting our data and for the data products we offer.