Chris Herring

I wrote the following for my current employers November newsletter, as insight to Data Visualisation.

3,892,179,868,480,350,000,000 or more simply put 3 sextillion, 892 quintillion, 179 quadrillion, 868 trillion, 480 billion, 350 million is the number of new digital information bits created in 2008 according to EMC Corporation. This figure calculates to 487 billion gigabytes, with EMC Corporation predicting that these figures will increase by five times in 2012.

The vast increase in the amount of data being generated, is down to the emergence of the “Internet of Things”, literally a device connected to the Internet or as Google’s Marissa Mayer defines as “Ubiquitous nano-sensors”. These devices are producing increasing amounts of data about the behaviour of the user interacting with the internet and the environment in which they’re doing so. So there’s a lot of data out there to visualise.

The process of creating a data visualisation starts with a question and meanders its way to construct a story that provides a clear answer to our question. Over the next few paragraphs I’ll briefly outline the 7 stages involved in creating a data visualisation.

Acquire
The first stage involves gathering some data and depending on your question this can be simple process or rather complex. Numerous public data sets are available for download on the web e.g. the Humane Genome (150 Gigabytes) or Weather measurements 1929-2009 (20 Gigabytes) are available from Amazon.

Parse
The data, which has been acquired, needs to be converted into its most useful format. This will involve tagging each individual piece of information for its intended use.

Filter
Once the data has been parsed, it will then need to be sorted so that any data not relevant is removed and/or normalised. This way all values are converted into a useful range.

Mine
The next step involves mining the data with math and statistics in order to extract a pattern that will form the basis of the story that will be visualised.

Represent
Within this stage the decision is made to the form that best suits the data. This could be a horizon graph, coloured circled placed on a map, or even a simple bar chart.

Refine
The decision made at the previous stage will influence refinement stage, where graphic design is added to help exemplify the changes in particular data. For instance, the coloured circles placed on a map could be transformed into particle clouds above the map, with denser areas representing areas of higher value.

Interact
The final stage adds interaction to the visualisation so that the user is able to control and explore the data. This can be as complex as filtering the data to create a new representation or just zooming in and out.

Here are a few interesting examples of data visualisation:
- Night vision maps of the WikiLeaks Iraq Casualty Data
- Cities from around the world traced using geotags and timestamps from Flickr and Picasa to determine the speed at which photographers traveled the landscape
- NetFlix rental patterns examined, neighbourhood-by-neighbourhood, in a various American cities
- The 372 songs sampled in All Day by the band Girl Talk visualised
- Flowing Data will visualise data about your environment, lifestyle and yourself using Twitter

Data visualisation has the ability to convey an insightful piece of information in an imaginative way, be in it print, web or even mobile. Marry this with a flashy or simple design and data visualisation can be an effective way to communicate a branded message.