Stop pretending data visualisation is easy – bring distributed skills together

I spent a great day at LocalGovCamp in Birmingham last Saturday, an unconference for anyone interested in how social media and digital technology relates to local authorities and improving public services.

LocalGovCamp Birmingham 18 June 2011
Photo by Glenn Wood

Toby Blume, of Urban Forum and Paul Evans ran a session on data visualisation and visualising policy (more on that in this previous post).

Partway through the discussion, one particular issue really grabbed my attention. There was some frustration from some local authority officers about how difficult it is to actually make a visualisation or to communicate issues visually. It went something like this:

“This stuff is really hard. I want a tool that will let me put my data in and will give me a nice visualisation back.”

After a few responses – useful suggestions such as starting with Google Spreadsheets or Fusion Tables – the frustration with the steep learning curve came out, and Michael Grimes refocused the room with this nugget of sense:

“The process of creation [for a data visualisation] is important. It’s about how we communicate accurately with the information we have.”

And Michael got me thinking… do local authority officers expect making a data visualisation to be a straightforward process? Should it be easier? Are the available tools not serving those new to visualisation?

Or, and this is my thinking, there’s a false expectation that visualising data is easy. The JFDI attitude prevalent in other areas of digital tools for local government may have created false expectations on ease of access to visualisation.

Other digital tools in the social web made for publishing content – such as free blog platforms, Twitter, Facebook Pages or sharing video on YouTube – are relatively straightforward to get started with and local authorities are using these tools to great effect already.

But ease of access to tools, having the ability to publish or the skills to find your way around a blog platform doesn’t necessarily mean you can communicate effectively. Also, you can’t learn to write well or communicate with other people by spending an afternoon reading blog posts on the subject. These are skills that take time to build up and are achieved through practice, experimentation and, frankly, well… work or experience.

Making an effective data visualisation of a civic issue or communicating policy ideas visually to help other people understand the issues is an involved process:

  • How should it be done?
  • Who decides what the message is?
  • Who ensures the data is accurate?
  • Who makes the visualisation?
  • Where is the visualisation shared for maximum effect?

The process of creating a visualisation requires a variety of skills:

  • Handling tabular data
  • Some understanding of statistics
  • Research and analysis
  • Understanding people, communicating ideas and storytelling with data
  • Visual design understanding and software knowledge

Let’s look at a few examples and try and dissect them a little.

How many households are like yours?

The New York Times published this interactive data visualisation which enables you to explore the different types of American households and see how these households have changed over time.

Screenshot: interactive graphic by New York Times visualising answers to the question how many households are like yours?
View full size image (png, 29kb) | Published by New York Times, 17 June 2011

This visualisation is a nice mix of charts, illustration and a simple user interface for users to select the household makeup to retrieve data on. Nice work – easy to understand – and an approach I could see being useful for UK local authority officers and elected members, particularly if similar UK demographic data at a more local level could be used.

Also note that four people were involved in creating this visualisation (Jeremy White, Ford Fessenden, Sergio Pecanha and Matthew Ericson).

Local apprenticeship data

Closer to home, the West Midlands Regional Observatory (disclaimer, my former employer) produced a set of visualisations on supply of and demand for apprenticeships in West Midlands local authority districts. The idea was to present the data in a single-page dashboard style for use by local authority chief executives and business leaders.

Screenshot: closeup of data visualisation of supply of and demand for apprenticeships in Birmingham local authority district

Birmingham apprenticeships dashboard by West Midlands Regional Observatory, January 2011

I’ve shown this project as I think it’s a nice example of what you can achieve with good communication and technical skills in Excel – skills readily available in local authorities. This visualisation was created using the tools the organisation had to hand – MS Excel and Publisher – with no extra graphics software.

Roundup and identifying skills

My response to the room at LocalGovCamp was:

Don’t beat yourself up with false expectations that you should be able to handle all of these areas on your own. It’s a rare person who combines all these skills. Graphic design, illustration, communication and statistical analysis are established long-standing disciplines. It’s unrealistic to expect yourself to quickly develop these skills together without a period of learning and practice.

Instead, find these skills distributed across a local authority in different people – and bring them together in an agile way when you want to create a visualisation.

Some more suggestions from the room are in Paul Evans’ slides, quoted here:

  • Review required skills for LocalGov employment
  • Co-ordinate visualisation skills within local government better
  • Lower expectations on corporate style – go for authenticity rather than branding
  • Encourage people other than formal employees to present information – it’s more authentic – enable and curate rather than ‘just create’
  • Make a clearer link between participation and decision making
  • Make organisations more permissive in comms terms – making everything go through the corporate filter doesn’t work
  • When we inform – say WHY we’re informing
  • Curate walk-throughs of how people do good data visualisation – dotgovlabs / skunkworks
  • Visual media surgeries!

A wizard-driven approach to visualisation

There was also talk in the room about a desire for a wizard-based tool to walk the user through creating a visualisation, with tips and recommendations suggested at relevant points in the creation process.

Have you used anything like this? Do you think such a tool would be useful?

Related links

10 comments

    • Moxy Park says:

      Totally agree with Gav and Mr John. Visualisations are an art form that require a human understanding of tabular data.

      And I think you’re right Gav, that people shouldn’t beat themselves up over finding difficulties. Building a compelling and meaningful website isn’t easy, nor is producing a visualisation that humanises tabular data. And just as there isn’t a wizard to build an imaginative website, I don’t think one would exist for visualisations that didn’t rely on templates and stencils taken from others’ original, sideways-on approaches to data (that’s not a value judgement on such a system, though).

  1. Andy Kirk says:

    Hi there, I was sent a link to this post so just wanted to introduce myself and share with you a few thoughts. My name is Andy Kirk, I am a visualisation designer and editor of the blog http://www.visualisingdata.com.
    This looks like a really interesting event and discussion you have had. The holy grail of a single tool/wizard for producing visualisations isn’t that viable, simply because, as has been mentioned above, it involves a lot of human input and iteration.
    My aim is to impart visualisation knowledge and skills to a broader audience, making this a much more accessible practice for many more people than traditional designers, statisticians or computer scientists. Increasingly we find ourselves facing greater challenges with taming and communicating data, even in roles not traditionally of an analytical persuasion.
    The aim of my site is to curate the best contemporary examples, practices, tools, debates and critiques to help people like yourselves develop confidence and capability within the field.
    You might like to check out an essential collection of resources I am (slowly!) compiling. They are captured in this summary post: http://www.visualisingdata.com/index.php/resources/
    There are around 50 of the most useful tools presented there (both free and paid for) with about another 30+ to come in the next week or two. I will be growing this resources list to include all sorts of invaluable references in due course.
    You may also be interested to know that, as well as undertaken visualisation design projects, I also offer customised training on visualisation practice – giving you the technical and theoretical means to become a capable practitioner yourself.
    If anybody would like to explore this subject matter more, just drop me a line – my contact details are on my site – or catch up with me on Twitter @visualisingdata.
    Thanks for reading!

  2. Gavin Wray says:

    Thanks for the feedback on the post.

    @Andy – interested to hear that you think that a single tool/wizard for producing visualisations isn’t that viable. I tend to agree with you at the moment, and think that there are too many human interpretations to go into a visualisation for such a tool to work. However, at the LocalGovCamp event there was certainly interest in a wizard approach. Maybe the human side of the technology isn’t quite there yet for such an idea to work, though that doesn’t necessarily mean it will never work as the tools and skills move on. (I loved using verifiable.com, which closed unfortunately in 2010, mainly because of its optional wizard-like tips walking you through data import, refining and labelling, on through choosing a chart types and publishing.)

    Also, like your list of visualisation resources. Very useful. I’ve bookmarked that.

    Cheers,
    Gavin

  3. Andy Kirk says:

    Hi Gavin, thanks for reply (and RT!). I see a visualisation design task as requiring both a suite and a chain of visualisation tools: The suite to fully cover the range of different approaches you might take (static/dynamic, standard > complex designs); the chain concerning the different stages you need to go through (eg. data exploration > visualisation production > final piece composition > presentation/dissemination). I’d suggest Tableau is the tool pushing ahead of the pack right now in terms of offering the greater range of principle-compliant visualisation creativity – Tableau Public is web based and free to use.

  4. Andy Kirk says:

    Just to add – I agree that we will get there one day with a single tool capable of almost everything, but that will only arrive once some of the key issues within this diverse field (convergence of art and science) are resolved irrefutably. Right now we are in an exciting era where creativity and innovation are pushing the field into new directions, and so it is evolving before our eyes.

  5. Gavin Wray says:

    Hi Andy. You’re welcome. Agree about the chain of stages involved and the different approaches. While there is overlap, each stage does require different skillsets and ability to choose relevant tools. Haven’t used Tableau Public for a while; will take another look. Good tip.

  6. Ching says:

    Interesting topic, but there is a growing number of tools that try to solve all problems in one (and claim not to have a learning curve). Usually that helps to minimise effort on either producing graphics/visualizing data, but the key missing component is the fact that all data tells a story.

    Sure, there can be a tool that does “input: data, output: visualization” automatically but it loses all meaning if it was difficult to comprehend or the pieces do not construct a whole picture. That said, I believe there can be tools to aid this process- not automate it to the degree of making the decision of how to piece it together.

Comments are closed.