Turning Numbers Into Pictures: The Science (and Certification) of Data Visualization

Numbers mean nothing to most human beings. This is why you can tell somebody that the likelihood of them winning the lottery is one in several million, and they'll shrug while buying the ticket and say, "Well, SOMEBODY has to win." Frustrated statisticians realized very early on that if data were going to be useful at all, then people would need to see statistical informaiton as more than just numbers.

Data visualization is the fancy term for "turning numbers into graphs," and it's actually a very valuable IT skill right now. It's equal parts art and science: A skilled data visualization expert must have both the mathematical ability to know whether data is significant, and the artistic ability to create visuals that can convey a message at a glance. Common visualizations include bar charts, pie charts, and timelines.

With the right program, however, almost any data set (no matter how obtuse) can be translated into an easily-understood graphic. Increasingly-sophisticated data-plotting software has made it so that data visualization doesn't need to be two-dimensional anymore, meaning that not only can you make sense of more variables than before, but also that the resulting visual looks much, much cooler.

While a skilled professional in almost any IT field could benefit from understanding data visualization, the discipline is especially useful in some niches. Because data visualization is a particularly valuable tool when it comes to identifying trends and patterns, Business Intelligence professionals tend to be masters of the practice. It's the BI professional's job to gather as much raw data on the market and competition as possible and process it to be useful to the decision makers of a company. Data visualization is essentially the only way to do that.

A little surprisingly, marketing departments are also commonly in need of data vis experts. Persuading a demographic to invest in a product or service becomes a lot easier if you've got strong but easily-digested statistics on your side. But the list hardly stops there. Web developers who have access to graphical apps can make the information on their site so much more palatable (in fact, there are quite a few sites dedicated solely to visualizing a particular set of data).

A well-designed mobile app that implements data vis will easily outperform apps with less accessible information. A network administrator trying to find weak points in a system can use a data plotting program to quickly pinpoint problems.

So, how does one develop these skills? Which programs should one use? Which degrees or certifications are best for it? Well, this is where things get tricky. Because data vis is not so much a position as a skill, what it entails depends largely upon the job role it's attached to.

Developing a good feel for graphic design and organization is a good place to start, however, and a basic understanding of statistics is always helpful. An ability to code will be useful if you're planning on cutting together your own plotting program for your own needs, and at the very least you must be willing to learn new programs as needed (so far, there don't seem to be many industry standards as far as visual data plotting programs go).

And as mentioned above, this practice is both logical and artistic, so if you're strictly one brain hemisphere or the other, then you're going to have to develope greater flexibility.

Because this is a fledgling discipline, you're going to want to learn from the experts — in this case, Edward Tufte and Stephen Few. Edward Tufte teaches a one-day course on the display of quantitative information and has written several books on the subject, and Stephen Few also teaches and runs a consulting company, Perceptual Edge.

Certifications related to data vis include graphic design certifications such as the Adobe CS5 or AutoDesk 3ds Max and graphical modeling certifications such as the SAS Certified Predictive Modeler.