Exploring data – a new approach to visualisation
“The purpose of visualization is insight, not pictures” – Ben Shneiderman
Today’s technologies make exploring data with visualisations an incredible valuable tool for gaining insights. Let’s examine what ‘data visualisation’ is, before focusing on ‘exploratory data visualisation’ and diving into some use cases to see how it transforms our everyday life.
Data visualisation in the time of Cholera
Over the past 150 years, different forms of data visualisation have become an essential part of academic dialogue. Graphic representations like maps or line, bar and pie charts are used to highlight speciﬁc data points and explain complex information in a simple and accessible way.
One of the best examples of a seminal early visualisation is this map (snippet), created by the British physician John Snow in 1854:
The map shows cases of cholera deaths in the Soho area of London with a bar for each case appearing at an individual address. More bars equals more cases of cholera. Adding the geographical dimension to the data enabled Snow to identify the source of the epidemic: contaminated water from the Broad Street Water pump.
With this graphic evidence, he was able to convince other scientists and doctors of the time of the link between the water system and the outbreak of the disease – and eventually revolutionise their approach to hygiene.
Explore and explain
The case above is a textbook example of a data visualisation: a visual representation of numerical data used to analyse and examine complex datasets and communicate their meanings.
Data visualisation commonly serves two goals: to explore and to explain.
- Exploratory visualizations help users analyse a dataset to uncover the patterns and trends. It is closely linked to statistical analysis – but with visual methods.
- Explanatory visualizations surface specific stories the data points towards, depending on the creator’s choice of view point. They are instruments for explaining and reasoning about real-life relations.
Even though hybrids of these types of visualisations are common, the construction of exploratory visualisations of datasets has up until recently been limited by the technical tools available.
Trends towards exploratory data visualisation
Recently, data visualisation has become a very active area of research and development again. Advances in technology are making the benefits of exploratory visualisation accessible to a broader audience.
A corresponding increase in both first-party, third-party and open data has led to a Cambrian explosion in visualisations: The widespread use of digital devices in all parts of life creates more data than ever and open data projects make this growing resource easily available.
Meanwhile, the availability of low-cost computing and digital tools make visualisation simpler for people, even without having a data science background. The popularity of applications like Tableau, R, D3 or Vega (just to name a few) help to make the visual exploration of huge datasets accessible to anyone with a computer.
As a result of the increased availability of datasets and easier access to visualisation tools, the ability of explore data becomes more democratic. It allows more people from various fields and industries to interact with data and gain new insights.
Let’s explore some use cases of exploratory and hybrid visualisations in real life:
1. Dashboard-based Monitoring
What gets measured, gets managed: Measurement is always the starting point of optimising operational processes. But data alone doesn’t guarantee insights. This is why more and more organisations are investing in dashboard visualisations to monitor their businesses.
The example above shows a dashboard developed for the Deutsche Bahn. It enables the railway company to inspect the masses of transport and passenger data that are gathered every day in great detail. Visual tools such as animated maps, stacked histograms and path-time-diagrams help identify peak times and manage the passenger loads more efficiently
A dashboard display of data can be useful in almost any organisation where big amounts of data from IoT sensors are involved. Visualisation help organisations react to problems immediately, but also discover patterns over time and inform decisions on where to improve.
2. Visual Models
Visual models can make physical environments accessible and enable people to explore them safely. These virtual representations are especially helpful for big-scale geographic and meteorological data or sensitive medical data.
Medical education is one of the fields, where there has been a significant boost in the amount of data in the last decades. Visualisation tools can give clarity, meaningfulness, and utility to the data.
The example above shows a digital autopsy. The high resolution and interactive visualisation is making sense of raw data from a CT scan. The visualisation aids criminal investigation without the use of a scalpel or any other invasive instrument. Ultimately, it increases the quality and efficiency of autopsy procedures.
3. Data journalism
Even until now, journalism tends to be primarily associated with explanatory visualisations. But like no other discipline, it has become the forefront in developing new ways to work with and communicate data.
Data journalism is a growing field that uncovers stories by analysing data. The Argentinian newsroom La Nación started its with a small team and limited budget.
For one of their visualisation projects they worked with volunteers to collect public data about the assets of statesmen and public officials from the scans like the one shown above. They digitised the information and turned it into a visual and interactive presentation, that allows users to search and read a large volume of information and do comparisons. Even though the possibility of exploring every original PDF document is still included, the digital interface is the key to understand the data and making it truly accessible to the public.
Not every form of data journalism does have to include visualisation. But the strength of interactive visualizations is that they can show findings by guiding the reader through datapoints of interest, but also encourage them to interact with the complete dataset. This changes the readers’ relationship with the data. It gives them more independence from the author’s bias, empowering them to ask their own questions and discover the answers in an intuitive way.
Data visualisation in the future
The ever-expanding breadth and depth of data has lead to a re-discovery of data visualisation. Moreover, the increased accessibility of visualisation tools and technologies has opened up the field for anyone, not just data scientists.
Explorative visualisation practices will be a huge part of this, enriching all sorts of industries with insights to inspire more considered and data-driven decisions. With data being the most valuable resource of our time, the role of visualisation will increase to make sense of it, unlock its potential and make its benefits more accessible.