The world produces a staggering amount of data every single day. Somewhere around 2.5 quintillion bytes. In fact, 90 per cent of all data was created in the last two years. We’re living on the vertical slopes of a hockey stick graph, where data production is rapidly outpacing physical storage (the so-called ‘Data Apocalypse’, where mankind literally runs out of server space, has scientists working on several strategies, like storing data inside DNA.)
This global data rush has led to another growth field: data visualisation. The more data you acquire, and more tangled the web, the harder it becomes to make sense of it all. That’s where visualisation comes in. Data visualisation is the act of taking raw numbers and putting them into an easily understandable visual medium, such as a map or graph.
Data visualisation examples
On some level, we’re all familiar with data visualisation. If you’ve ever used Microsoft Excel to produce a bar graph or a pie chart (or even a simple table), that’s data visualisation in action. But the field has evolved alongside new technology over the years, which means data scientists now have dozens and dozens of options when it comes to making data look pretty: bubble clouds, heat maps, radial trees, mekko charts. Anything that can turn numbers into narrative. “Numbers have an important story to tell,” says data visualisation expert Stephen Few. “They rely on you to give them a clear and convincing voice.”
Why data visualisation matters
You might think, well, is data visualisation really necessary? Especially with the rapid job growth in data science. Surely highly trained data experts don’t need colourful pie charts to make sense of numbers?
There are a couple of problems with that assumption. The first is the sheer scale and complexity of modern-day data: in many cases, Big Data sets are simply too vast and complex for traditional processing software, let alone the human brain, to manage effectively. This is partly why AI and machine learning have caused big waves in data science over the last three years.
The second problem is the assumption that data exists only for the IT department, rather than as a business-wide resource. Data visualisation is one of the final (and most important steps) in Joe Blitzstein’s famous data science process: after data is collected, processed and modelled, the relationships need to be visualised. Not just so that data scientists can draw conclusions, but so they can sell those conclusions to the wider business. Data visualisation makes data approachable – and more importantly, useful – for everybody. Like Stephen Few said, numbers don’t mean much until you give them a story.
Picking out trends
Data visualisation also improves the quality of overall insight. Put simply, it does things that traditional descriptive statistics (ie. pages and pages of raw numbers) can’t. The most famous practical example is the Anscombe’s Quartet, devised by statistician Francis Anscombe in 1973: four graphs that have nearly identical descriptive statistics, but totally different visual distributions. It might take a data scientist weeks of trawling through spreadsheets to find this kind of pattern, but thanks to data visualisation, we can spot outliers, anomalies and trends almost instantly. This obviously saves companies time and money, and improves overall decision making.
Data visualisation vs Information visualisation
You might hear the phrase ‘information visualisation’ alongside data visualisation. The two terms aren’t quite the same thing. Broadly speaking, information equals data plus meaning. Information visualisations usually take the form of infographics, specifically designed to communicated information in a certain way. Data visualisations do the same thing, but they’re generated directly from raw statistical information – they’re literally the numbers in visual form. Information visualisation is less prescriptive. It’s more concerned with getting a message across, and usually involves data that’s already being processed. If you need a simple example, think of the difference between an illustrated infographic, which involves deliberate visual decisions, and a computer generated bar chart.
Jobs in data visualisation
As you might expect, the market for data visualisation experts is excellent, and expected to remain so. It’s hard to think of any industry that couldn’t benefit from making Big Data more understandable. Media outlets like the New York Times now have dedicated teams to produce interactive graphics. Analytics groups, data research labs and other, in-house visualisation positions are also growing rapidly. In some ways, with the rise of ‘citizen data scientists’ and AI-driven automation, data visualisation skills are more important than ever. Businesses need people who can walk in two worlds now: data analysis and visual storytelling. If there’s a more lucrative section of the digital Venn Diagram, it hasn’t been discovered yet.
Want to learn more about data visualisation? Check our RMIT Online’s data visualisation short course, powered by Udacity.