Data visualization is a tool, too!

I think everyone would agree that data visualization is all about making information accessible. But after some discussions at the see conference #4, I’m not sure whether everyone agrees about the target audience of data visualizations: A common conception is that data visualizations should be understandable by everyone because – in contrast to a multi-page scientific text – it is their purpose to be easily readable. It is my feeling that this thinking limits the potential of data visualization as a tool. Let me explore this with an example:

brain-tools-1

The above visualization of a brain showing where and when neurons fire is certainly readable to everyone who knows what a brain looks like, but in no way does it convey the meaning of the data to anyone but experts on brain science. It doesn’t even have labels that would clear things up. It is a tool to explore and think about data visually, not the end result of the research. Interestingly, nobody argues that because it is “scientific”.

But some people seem to think, that a visualization coming from a designer should be immediately graspable because it is “beautiful” and “designed” and therefore the design fails if they don’t get it. We think so ourselves sometimes, always wanting to create simple things that even our mothers can use. Simple is great, of course, it’s what motivates us! It’s just that if we are designing an expert tool, it may very well require expert knowledge. If we are clear about our target audience and their abilities, we will be able to produce more powerful tools by building on their knowledge and expanding it, instead of restricting ourselves thinking we have to keep things simple.

So, who is using the tool and how are they using it. Is it about consuming data, or working on data. Is it about creating knowledge, or taking it up? Are we building a tool for consumers or a tool for experts? Fill in your own thinking here …

Data visualization is an important way to think about data. Even though it is often used to “just” visualize some numbers to make people think about a topic, we should not neglect it’s power as a tool to work with data and create relevant knowledge.

Share this article

Subscribe for more

Give Feedback

  • http://herrstucki.ch/ Jeremy

    Very nice post!

    What bothers me sometimes is that data visualization isn’t seen as a useful tool but rather as a “nice-to-have” design feature that’s just supposed to present pretty pictures (I’m talking about clients here). I guess, it’s not only our job to consider expert users but to also show them that it is really a powerful tool.

  • http://herrstucki.ch Jeremy

    Very nice post!

    What bothers me sometimes is that data visualization isn’t seen as a useful tool but rather as a “nice-to-have” design feature that’s just supposed to present pretty pictures (I’m talking about clients here). I guess, it’s not only our job to consider expert users but to also show them that it is really a powerful tool.

  • http://saaientist.blogspot.com jandot

    Very valid point, Peter. As a researcher working in genomics and handling big datasets, I really believe there is an important place for visualization in knowlegde discovery. Statistics is the main tool to investigate datasets, but it requires assumptions before you can start analysis. This can influence your results. The best tool to find patterns however are your eyes and brain, not a statistical analysis. So my take on that is: first try to visualize research data as raw as possible (without pre-analytic steps), and only _then_ start any statistical analysis. This is obviously not relevant for all types of research, but it should nevertheless be kept at the back of our heads.

    Wrote a brief blog post about it on http://saaientist.blogspot.com/2008/11/visualize-or-summarize.html

  • http://saaientist.blogspot.com Jan Aerts

    Very valid point, Peter. As a researcher working in genomics and handling big datasets, I really believe there is an important place for visualization in knowlegde discovery. Statistics is the main tool to investigate datasets, but it requires assumptions before you can start analysis. This can influence your results. The best tool to find patterns however are your eyes and brain, not a statistical analysis. So my take on that is: first try to visualize research data as raw as possible (without pre-analytic steps), and only _then_ start any statistical analysis. This is obviously not relevant for all types of research, but it should nevertheless be kept at the back of our heads.

    Wrote a brief blog post about it on http://saaientist.blogspot.com/2008/11/visualize-or-summarize.html

  • http://www.behance.net/LucaMasud Luca Masud

    Great post! We (at densitydesign.org) were discussing of this just some days ago. It looks like this is not well understood even in the academic circle, but we are struggling to promote this view. :)

  • http://www.behance.net/LucaMasud Luca Masud

    Great post! We (at densitydesign.org) were discussing of this just some days ago. It looks like this is not well understood even in the academic circle, but we are struggling to promote this view. :)

  • Pingback: Help us refine: Categories on Datavisualization.ch

  • Pingback: Is Data Visualization a way of story telling, exploration, or expression? « design - code - interaction

  • http://marcinignac.com/ Marcin Ignac

    Hi! Recently I was thinking about the same issue. Is Data Visualization a way of story telling, exploration, or expression? Depending on which of this aspects you want to focus on it can be anything from a tool to ‘infoporn’. And of course nobody forbids you from mixing some of this ways together. Especially telling stories but being able to go deeper and explore data underneath the story by your self is something that I observe happening more and more in interactive info graphics on the web now.

    More on my blog: http://marcinignac.com/blog/2009/06/15/is-data-visualization-a-way-of-story-telling-exploration-or-expression/

  • http://marcinignac.com/ Marcin Ignac

    Hi! Recently I was thinking about the same issue. Is Data Visualization a way of story telling, exploration, or expression? Depending on which of this aspects you want to focus on it can be anything from a tool to ‘infoporn’. And of course nobody forbids you from mixing some of this ways together. Especially telling stories but being able to go deeper and explore data underneath the story by your self is something that I observe happening more and more in interactive info graphics on the web now.

    More on my blog: http://marcinignac.com/blog/2009/06/15/is-data-visualization-a-way-of-story-telling-exploration-or-expression/

  • travc

    Preach it brother ;)

    As one of those ‘scientists’, I can attest that visualization is not accepted to support results very often. As Jan points out, statistics is the primary tool. That said, visualization is critical in exploring the data and more and more frequently, explaining it and why the statistical analysis makes sense. This isn’t fair of course, since many types of visualizations *are* presenting the results of statistical tests.

    I’d also like to complain a bit about the reluctance amoung scientists to accept new visualization techniques. It makes sense, familiar methods are more trusted and the readers have trained their brains to understand them at a glance. Still, even very simple changes to objectively better and more informative presentations are resisted. Not a ‘visualization’ in the narrow sense, but every time I see a box plot I want to throw something.

  • travc

    Preach it brother ;)

    As one of those ‘scientists’, I can attest that visualization is not accepted to support results very often. As Jan points out, statistics is the primary tool. That said, visualization is critical in exploring the data and more and more frequently, explaining it and why the statistical analysis makes sense. This isn’t fair of course, since many types of visualizations *are* presenting the results of statistical tests.

    I’d also like to complain a bit about the reluctance amoung scientists to accept new visualization techniques. It makes sense, familiar methods are more trusted and the readers have trained their brains to understand them at a glance. Still, even very simple changes to objectively better and more informative presentations are resisted. Not a ‘visualization’ in the narrow sense, but every time I see a box plot I want to throw something.

  • Pingback: Scientific Visualizations on TED on Datavisualization.ch

  • http://www.flexmonster.com/ Arturas

    I believe that the perfect would to get both, but sadly we don not live in a perfect world ))) Though i believe that the trend for customizable dashboards and charts will last for a long time

  • http://www.flexmonster.com Arturas

    I believe that the perfect would to get both, but sadly we don not live in a perfect world ))) Though i believe that the trend for customizable dashboards and charts will last for a long time

  • Pingback: Shaking the “Pretty Picture” Stigma « WebGlide

  • Pingback: Shaking the “Pretty Picture” Stigma « WebGlide - Data-Visualization

  • Pingback: Data kan ook niet saai zijn! |

  • Pingback: Shaking the “Pretty Picture” Stigma | eagereyes