Visualising helps make sense of data
Teresa Tocewicz takes a look at how visualisation can be used most effectively for data analysis
Big data, big problem? When it comes to data, we’re getting more and more of it – but our ability to understand that information is far more important. At Jisc we’ve run a competition inviting people to visualise library data, as Google and the Guardian have done with open government data, and the British Library and others are now doing the same for the key data from the seven UK research councils.
With the advent of online tools, creating visualisations has never been easier – or so it appears. The 2011 McKinsey Global Institute report famously declared that by 2018, the USA would be short of between 140,000 to 190,000 employees with ‘deep analytics capacity,’ and an even greater scarcity of ‘data-savvy managers (1.5 million).’ Nevertheless, managers in the knowledge industry are turning with ever greater interest to their figures to justify everything from e-resource spend to research impact. Even without specialised analytics training, being able to visualise their data accurately has never been more important.
For researchers and knowledge workers, the challenge is not only to make their work accessible to others, but also to use their charts and diagrams as a way of seeing things differently. By replacing information with image, we can often see a different story hidden in the data.
The best visualisations are instinctive and imaginative. They are not over-complicated by design and they point only to the main messages. Using shape, size, colour and position clearly, numbers become ideas. Through them we see interesting relationships which we wouldn’t appreciate from tables full of figures. We can then interpret these representations. It is faster and easier for the human brain to analyse images than tables – patterns can be spotted quickly and easily.
Potential pitfalls
Magazines and newspapers are full of bright, colourful graphics which draw us into the stories. Their infographics show trends of sometimes global proportions, often using massive datasets.
But in the education, museum and archive sectors, we’re not always in that position. Universities, for example, could be dealing with relatively small numbers of postgraduates working on a particular course. The challenge for the person visualising that data is not to over-simplify, or encourage people to draw conclusions based on minimal information.
Moreover, data analysis is all about relationships between things, so context is of crucial importance. Providing these crucial details is the difference between art and analysis. As Bitsy Bentley, director of data visualization at GfK Custom Research, has said, ‘It’s difficult to balance the “wow”of a visually striking presentation (that tends to sell well), with the ‘aha’ moment when the user is able to easily and simply understand the problem at hand.’
Getting it right
Visualising data successfully requires time and training. For those interested in finding out more about what to do, and why, the Jisc infoKit on data visualisation can help. For example, with so many data presentation choices available, it is sometimes difficult to know which type of chart to choose. Several websites can help, such as Graph Selection Matrix, which presents your options in a walkthrough diagram and Nuts and Bolts & Graph Types (infographics), which provides a light hearted but informative summary of charts.
Identifying the audience need is also crucial. For an audience new to graphs, annotations are helpful. Allowing the user to add new data items or see the information in different charts can also make the visualisations easier to handle. For a more skilled audience, a more innovative approach can help draw attention to those data stories. If the charts are accessed through the user’s mobile device, then they’ll need to be scalable.
Future opportunities
Our age of big data, and small data, provides us with so many opportunities. According to the Pitney Bowes white paper on Location Intelligence: The New Geography of Business, more than 80 per cent of all data maintained by an organisation has a location component. Thanks to advances in technology, we are now in a position to add that geo-location data to a visualisation – for example, to allow someone looking at a map of population statistics to see only the data relevant to the area they’re standing in. Mash-ups of data from disparate sources will all provide insights in the future that we can only dream of. In our view, it will take accurate visualisations to make those dreams a reality.
Top 10 tips for data visualisation
1. Keep it simple – avoid ‘chartjunk’ at all costs;
2. Focus on purpose – don’t try to do everything with one visualisation;
3. Choose the right chart - understand the strengths and limits of each chart type;
4. Make use of micro and macro perspectives - think about your audience: are they senior management or practitioners?;
5. Avoid information overload - multiple charts are better than cramming information into one;
6. Avoid 3D charts - the perspective can distort the data because what is displayed at the front is perceived as more important than the background;
7. Amplify the message - use the title to emphasise the main finding in a compelling headline
8. Don't confuse visualisation with data art – the main job of data visualisation is to inform and help someone understand the data; the primary purpose of data art is to deliver an aesthetically pleasing experience and to entertain;
9. Proofread - often we focus so much on the numbers that we forget about the words, their spelling, grammar and clarity of the message;
10. Use colour sparingly - colours in analytics should add meaning and extra context to the visualisation, not just to make the charts look pretty.
Teresa Tocewicz is a researcher and analyst at Jisc infoNet