As you may just have heard (we’ve been a tad over-excited…) data visualisation maestro Aaron Koblin came into to talk to us yesterday. He kicked off with a showcase of his work, from his exquisite grad school visualisations of flight paths (see post below) to his latest embryonic projects for Google labs. Along the way he showcased extraordinary visualisations of the ebb and flow of information in cities and around the globe, experiments in crowdsourced sound design and perhaps his most famous project, the Radiohead “House of Cards” promo.
In showcasing his extraordinary portfolio he touched on a number of powerful and provocative themes which we followed up on in our interview. Themes around the power of social context to make data compelling, the power of data visualisation to embrace the complexity of our lives today and the tension between the human and the machine present in crowd-sourcing engines. He also shared his key learnings from life at the front line of data visualisation:
Looking at everyday things in new ways completely changes your perspective: there is no “mundane” data when you set it in context.
Use multiple visualisation techniques: there’s more than one way of seeing things
Stay true to the data, not the “real world” : There is a random-ness to data-it will make patterns you never anticipated. Respect the random-ness.
You don’t have to use all the data : sometimes seeing patterns is about what you leave out
Set the data free: open-source and let other people play with your data
Following his talk, very graciously agreed to be interviewed by Labs about our (and your) burning questions around data visualisation. It was a fascinating conversation for us and we hope for you. So over to Aaron….and many thanks to those who submitted questions for him.
Why do you think the world has suddenly gone crazy for data visualisation? 18 months ago it was a struggle to get anyone interested in data and now it’s the new rock and roll…
I guess it’s really the times that we live in, now you have tools like Twitter and Facebook and things that are widely used not just by the nerds but by everybody. Popular culture has also just all of a sudden embraced the power of storytelling through data and the relevance of all the data to their lives. All kinds of things have happened that simply weren’t possible before – the author you look up to, the musician, etc. they’re sharing all kinds of things – you can be intimately living their lives along with them and you see all different types of applications.
Do you think it’s partly about the explosion in the amount of data currently available, the data trail we leave behind us now or the fact that companies have more data than they can process so they end up giving it away?
I think ultimately the biggest change is that the data is now relevant to people’s lives. Before most of the data was about infrastructure at best and a lot of it was locked away or presented in aggregate form. When you’re presented with a huge lump sum number that has no context it’s just not interesting, but now when you get these granular stories, things that are saying at this specific point in time here’s the way that things changed, just by giving it that context and social relevance it becomes interesting.
Perhaps the big difference between what you do and the bar chart or the single number that it really embraces complexity rather than trying to reduce everything-ur lives are complex and this gives you a deeper understanding of that, not simpler, but richer
I think what’s really nice is when you can have that kind of simplicity but then allow for investigation. Not necessarily forcing it into this sterile reality, but being able to present a story clearly and convincingly and simply but then allow for justification where you can say this is why, it’s fine to give a summary number but then be able to say this is why the number exists.
So do you think data visualisation should be about immediacy or intrigue? Should it be “I see that and I get it” or “I see that and I don’t get it so now I’m going to play with it”?
Definitely it boils down to your purpose. I think there certainly is a place for scientific visualisation. There still is a necessity for that type of clarity and objectivity but there’s also a place for design and I would argue for art, that to be able to use data to tell stories and to tell the right types of story requires different kinds of techniques and different means. Often times I think the whole medium-is-the-message sense of actually using the system to think about the system can be valuable and fun and productive
What do you think is so fascinating about seeing one medium like music or dance portrayed through another?
I think that’s also something that’s really picking up because of digital culture. Now that everything has become digital it’s so easy to run it through a completely different process. You can make it, just tweak the algorithm and sound becomes image and image becomes motion. It’s kind of a natural process, it makes a lot of sense, especially for people like myself who are visual thinkers and learners. I think translating a lot of these concepts and numbers and pure abstractions into something tangible, something to be seen and experienced and interacted with means the world, because for me it makes a completely different kind of sense. A lot of times that kind of experiment can reveal the underlying structure and point out the way that it makes sense.
You talked letting the data do the talking and really embracing the random-ness of the data; do you think that’s what makes data visualisation so compelling as art, because art is very seldom truly random?
I think it really gives character, because I think it’s really that kind of intricacy and detail that builds character and in a sense it’s the errors and flaws that make art. If you look at creative practice – like with the Sheep Market project for instance, if every person drew a perfect sheep they would all be the same and it would be a horrible project. It’s actually seeing the ways that people fail, the different intricacies and character that comes from the individual that adds a lot. You see that in all data visualisation, it’s the little variations that gives the character and makes it interesting.
On the crowdsourcing sheep project, you talked about it potentially being a very fragmented and alienating experience but you’ve drawn it into a coherent whole-there’s a certain ambivalence there
I think it was the juxtaposition of those two qualities that makes it interesting for me. On the one hand you look at this huge grid that looks very much like a matrix of computer created content, but then juxtaposing it with each individual, looking down at the fine level you see there are actual little people in there. I’ve always been interested in microscopes, I bought a few microscopes and I have a family friend, Gary Greenberg, who makes these amazing oblique-lighting microscopes. Basically microscopes that produce images that are more like beautiful photographs than back-lit medical tools. He used to let me play with them and it was amazing fun. This notion that there’s a device that can completely transform the way you see something is really inspiring. You can look at the whole thing but you can also drill down and it’s a totally different world.
The wisdom of the crowd versus crowd-sourcing is a fascinating topic. The wisdom of the crowd seems to kick in when the crowd doesn’t know it’s being watched, whereas with crowd sourcing it can sometimes just feel like mass-sourcing. Do you think the future of crowd-sourcing is genuinely collaborative, with the crowd consciously working together and making things better?
I think you already see that happening with all the Wiki projects which are really inspiring. To some extent I think it’s because the motivation is different, it’s not really about money. Money really complicates things. With the sheep the people who I had paid two cents felt totally ripped off and were really mad at me but the people who knew what the project was were asking if they could give me free sheep! So they had a completely different perspective on the situation which was interesting to note. I think the weird thing about crowdsourcing is that to some extent it feels like the inevitable evolution of capitalism which is just something to think about.
Which is weird, because on the other hand you could argue it’s socialism in action….
Right, it’s both – it boils down largely to the approach of working with the crowd. There probably needs to be a better disambiguation for actions involving the masses. There’s wiki-style collaboration, the kind of thing you see with open source projects, people working together to make something massive. On the opposite end of the spectrum is the current incarnation of the Mechanical Turk, where you have individuals being harvested for isolated menial tasks, and somewhere more towards the middle you find things like CrowdSpring – more of a massive sifting of the crowds.
There are some themes which seems to recur in your work, such as the energy of cities or some themes about the flow of information. Is that about themes that interest you or is it about the data sets that are readily available?
I wouldn’t yet say readily available, it’s still really tough to get at some of that data, but I think I am generally interested in all kinds of data that have anything to do with our lives and revealing the way that we live and build systems. So to me it’s something I’ve always been drawn to and part of it obviously is because I grew up in this computer and game culture and I’ve always been interested in algorithms and the way things work and the process behind things. There were a lot of films growing up about information and visualising information that inspired me a lot.
What tends to come first for you, the data set or the visualisation technique?
Based on the projects that I’ve done it’s usually either that I’m presented with an awesome dataset or it’s that there’s a data set I’d really like to create. I guess that’s answering it by saying both. So with the mechanical turk projects it was more about being interested in that tool and wanting to create data that would reflect the tool.
Does the software you use have a big impact on the way you work? You talked about the impact of the Processing tool?
I think it definitely has. The nice thing about Processing is that it’s an open source tool so it’s constantly being added to and growing and because it’s open source it makes it much easier not to be heavily influenced by it. I think because you can modify it to an extent that’s much more thorough than if you were using a closed source tool. In the sense that there’s all kind of things people have written that you can use but also in the sense that if something doesn’t work the way you want it to you can rip it apart and make it different and make it work the way you want.
We are living in a golden age of data availability right now. Does it bother you either that there is such a rich data trail available about our lives or that people may start withholding data on that basis?
I think that we will see people change, at least in terms of personal data, I think we’ll see people change their interest in sharing as much as they are. But I think that will probably come in the form of not necessarily less data acquisition just less public data sharing. I think what we’ll probably see is better disambiguation between aggregated public data and individual public data where I’ll be willing to opt in to something to share my information but not with my name on it and I think that that will end up being really valuable for all kinds of social studies and applications. But I think that will also potentially be less damaging to individuals as we see more of that. I feel bad for the younger generation that’s growing up right now. A lot of the stories that they’re bonding to their existence, will leave trails that will last with them for the rest of their lives. Forgetting is a wonderful ability, and one that technology is not currently adapted to.