Posted on 14 July 2015 - By Helen Kennedy

How people engage with data visualisations and why it matters

Different types of chart:

As data visualisations become increasingly ubiquitous, just like the data on which they are based, the ability to make sense of and engage with them becomes increasingly important. 

In this context, visualisation literacy is a necessary skill for living in our information and (big) data times. Data visualisation expert Andy Kirk of Visualising Data estimates that there are around 75 common chart types, and that’s just the ones that have names.

The proliferation of visualisations and chart types can make understanding them difficult, yet it is also essential for those of us who wish to make sense of the data within them and so to participate in informed ways in data-driven conversations and society.

Previous research

Until recently, not much was known about how people engage with data visualisations. A lot of visualisation research and literature focused on how to design effective visualisations, but much of this was based on authors’ design instincts, rather than research with users, such as the influential work of Edward Tufte.

The limited research which has involved users often pre-defines effectiveness in very narrow ways, such as memorability of data, speed of comprehension or task completion, and it then proceeds to measure these things. Such research almost always focuses on visualisations themselves, looking at things like Tufte’s famous data:ink ratio, chart junk, visual metaphors and grouping of elements.

Research into engagements with visualisations almost never focus on the people doing the engaging, the skills they needed to do so, or how the social contexts of engagements affect the process. 

The Seeing Data project

This gap in knowledge and approaches led me and a team of researchers to examine how people engage with visualisations on a research project called Seeing Data.Through diary-keeping and focus group research (with 46 participants), long-term diary-keeping and interviews (with 7 participants), interviews with (13) visualisation designers and other methods, we identified a range of factors which affect people’s engagements with data visualisations. These include: 

  • Subject matter: when the subject matter speaks to our interests, we’re more likely to engage with data visualisations.
  • Source/location: when visualisations are encountered in already-trusted media that we view or read regularly, we are more likely to trust them; otherwise, we don’t.

"I watch quite a lot of the news and I think sometimes they use graphs and charts and things to highlight their issue." (Sarah, 34, retail assistant)

"To confuse you." (Chris, 38, agricultural worker)

  • Beliefs and opinions: we like visualisations which communicate data in a way that fits with our world views, but some of us also like our beliefs to be challenged by data and visualisations.
  • Time: engaging with visualisations can be seen as hard work by people for whom doing so does not come easily, so having time available is important in determining whether we want to do this ‘work’.

"Because I don’t have a lot of time to like read things and what have you, so if it’s kept simple and easy to read, then I’m more likely to be interested in it and reading it all and, and you know, to look at it, have a good look at it really" (J.C., 24, agricultural worker/engineer)

  • Visual elements: the conventions that visualisation designers draw upon play a role in determining whether we’re willing to spend time looking at a visualisation. These may appeal to us, or they seem too unfamiliar and therefore offputting.

"It was all these circles and colours and I thought that looks like a bit of hard work; don’t know if I understand." (Sara, 45, a part-time careers advisor).

  • Emotions: visualisations provoke emotional reactions – if we feel immediate confusion about a visualisation, we are less likely to invest time and effort in making sense of it. Subject matter, visual style and other factors all provoke emotional reactions.
  • Confidence and skills: we need to feel confident in our ability to make sense of a visualisation, in order to be willing to give it a go. This usually means feeling like we have some of the requisite skills.

These factors haven’t figured prominently in research into engagements with visualisations, so our findings represent an important contribution to understanding how people engage with visualisations and the factors that affect this process. Taking these factors into account is vital in order to fully understand user engagement with visualisations.

Visualisation literacy skills

The last point, confidence and skills, is especially pertinent to people involved in information literacy education. The skills that we and our participants identified as integral to ‘visualisation literacy’ include:

  • Language skills, to be able to read the text within visualisations (not always easy for people for whom English is not their first language).
  • A combination of mathematical or statistical skills (knowing how to read particular chart types or what the scales mean) and visual literacy skills (understanding meanings attached to the visual elements of datavis) – sometimes called ‘graphicacy’ skills.
  • Computer skills, to know how to interact with a visualisation on screen, where to input text, and so on.
  • Critical thinking skills, to be able to ask ourselves what has been left out of a visualisation, or what point of view is being prioritised.

Improving understanding of data visualisations

To begin the process of developing these skills and visualisation literacy more generally, on Seeing Data we developed a resource targeted at non-experts, called ‘Understanding Data Visualisations’, which is made up of a mix of quick-and-easy, interactive and in-depth content. There are very few resources like this. The Data Visualisation Catalogue, The Graph Design IQ Test, and The Visual Quantitative Literacy Test are exceptions, but these test more than they educate.

Our resource is just a start, and much more is needed – more resources, more discussion of the skills needed to engage with visualisations (aka visualisation literacy) and more understanding of how best to undertake the project of developing visualisation literacy.What are the best ways to do this? Traditional methods, interactive learning,co-creation, hacking? How does visualisation literacy relate to statistical literacy and data literacy? How can we ensure that people develop skills in digital and visual literacy, in looking as well as reading, in being able to make sense of the multimodality of visualisations and in understanding them in social context, as social constructions which hide as much as they reveal, communicate a particular point of view, or choose not to tell particular stories about the data on which they are based? 

Some of the findings of our research raises questions about how people learn to relate to data, through formal mathematical education and more informal cultural practices. For example, we found that our participants related to statistics emotionally as well as cognitively and rationally – emotions is one of the factors that affect engagement, and our research data is replete with quotes in which participants express feelings about visualisations themselves, the data being presented, the subject matter or the sourve of the visualisation.

So, we might need to re-think existing, scientific approaches to statistical education and consider what arts-based approaches might contribute to developing skills for feeling confident with data.This is an important point, because changing how we do statistical education might mean that participation in our increasingly data-driven world becomes more open to more people.


Image source: Seeing Data website

Related knowledge and skills


Read our blog comment guidelines