I’m in California with Brian and his parents! The trip has been great so far and I’ll post a bit about it when I get home. In the meantime, I’ll share what I was up to last week: the SXSW Eco conference.
SXSW is well known as a big conference that meets in Austin every year in the spring. There are several aspects to it: music, film, and interactive. The Music festival brings word class talent to Austin, contributing to our reputation as the Live Music Capital of the World. The film festival is probably the least well known amongst the general public, but it’s still considered the place to be for celebrities that week. And the interactive festival is renowned for attracting the up-and-comers of the tech world and being the platform of choice for publicizing new startups that have later gone on to fame and fortune, like Twitter.
Recently, the SXSW enterprise has started sponsoring a few other conferences, like SXSWedu for education and the one I attended, SXSW Eco, covering environmental innovation and design.
SXSW Eco started several years ago in the fall and I attended the inaugural conference thanks to free tickets I scored by being a climate scientist at UT. The experience was a great one. It was a different (non-scientific) crowd than I usually get to interact with at conferences and a lot of the topics were interesting and inspiring.
This year, I got a similar opportunity to attend and I jumped at it. Unfortunately, I had a lot of other meetings and things going on at work so I wasn’t able to attend a lot of the sessions I was interested in, but I sat in on a few which gave me a reminder of the perspective of the environmental community.
As a climate scientist, I consider myself somewhat outside the environmental movement. Environmentalists are often pegged as being activists or hippies, which I don’t think is necessarily the case. But by and large they focus on igniting change in the status quo to protect the environment and develop sustainable societies. As a scientist, I study climate and how the earth is changing, but I make no suggestions about what society’s response should be to those changes. I help demonstrate the possible impacts of climate change and the risk they pose to society, but my work is meant as reference for other people making decisions about how to proceed, rather than an outline of what should be done.
While most climate scientists don’t pair their science with moral implications, there’s an inevitablel undercurrent of urgency for action as we learn more about what’s happening to the earth. That’s because with the physics and chemistry and biology of the earth comes the statistics, which are warning us that the risk of significant climate impacts are substantial. That leads to strong implications of future social, economic, and political ramifications for not acting or waiting too long to act to mitigate or adapt to the earth’s changing climate.
SXSW Eco is, in part, a place for people to come together to share creative ideas about educating the public about climate change and its impacts and on directing social change that will reduce the pain inflicted by these impacts. The presentations that most interested me had to do with the communication of science as evidence for the seriousness of the threat that climate change poses.
One of the talks I attended had to do with data visualization. It was presented by a representative from a design studio called Stamen. Stamen does consulting work with research groups who are seeking better visualization for their data, often for the purpose of being able to explain their results to the public. She gave several examples of their work, but one in particular hit home for me: mapping subaquatic vegetation (SAVs) in the Chesapeake Bay.
The Chesapeake Bay ecosystem is fragile, largely as a result of pollution. SAVs help to improve the water quality, control erosion, and sustain wildlife, but they’ve been on the decline for years. When I was growing up near the Chesapeake, elementary students took field trips annually to help replant SAVs. Local scientists would demonstrate the water quality by basically sticking a ruler down in the water and recording how far down you could see; water clarity was a proxy for water quality.
Stamen did a project to help visualize the distribution of existing SAVs in the Chesapeake and to track their progress over time. (Note that the website is currently unavailable due to the government shutdown; hopefully it will be back up soon.) They got their data from large databases collected by scientists and were advised by experts along the way. And George Oates, Art Director at Stamen, had some tips for how to best visualize scientific data.
They say that beauty is in the eye of the beholder and that applies to science as well. For example, mathematical beauty is an actual thing. It has its own wikipedia page, so you know it’s legit. “Beauty” in math and science often boils down to simplicity and elegance. Simple, straightforward solutions are considered most “beautiful,” even if it takes scientists and mathematicians years of work to get to the point of distilling their results into such a thing of beauty.
An example of this I work with frequently is the Laws of Fluid Dynamics. There are 3 laws: mass, momentum, and energy must be conserved. All aspects of the earth system rely on fluid dynamics: we think of the atmosphere, ocean, ice, and even the earth’s mantle to be fluids. When you can describe all that with 3 simple rules, expressed in 7 words, that’s what we call elegance in the scientific sense.
Scientists need to remember that the rest of the world doesn’t think the same way, and so its necessary to visualize data products in a more traditionally beautiful and simplistic way that conveys the main points to any audience.
Speaking of what’s obvious or not, we need to cut to the point. Too often when trying to communicate science, we take too much for granted. I’ve agonized over whether I’m dumbing things down too much for an audience or whether I should give more background information.
But the fact is, I’ve rarely presented or listened to a talk that gave too much basic information to introduce a topic. It’s not always the case that the audience is seeing the material for the first time, but giving the science some context allows an audience to be immersed in a topic they don’t deal with all the time. Scientists shouldn’t be afraid to state the obvious and to do so simply. As Einstein said, “if you can’t explain it simply, you don’t understand it well enough.”
Stamen specializes in producing spatial maps to demonstrate data. This kind of approach doesn’t apply to all science, but using maps in both time and space can go a far way to tell the story behind the science. For this reason, it’s often a good idea to use animations or video to communicate science.
This can be one of the trickiest things to do for scientists and is therefore one of the best places for designers to help us out. Scientists like to work with graphs. We likes plots, lines of best fit, error bars, axes. We’re used to seeing these and analyzing what they mean. The general public is not. While doing analysis, I’m often asking myself, “what plot could I make that might reveal something about what this data is telling me?” But there’s a much larger world of data visualization out there that I don’t approach, mainly because I don’t have the skills to make it happen.
The photo above is a good example of this. The slide shows visualization of wind patterns in the U.S. This was a personal project done by a couple of designers using publicly available data. The site spread quickly on social media. At first glance to a scientist, this kind of visualization may not seem like it contains enough information. The wind speeds are generalized, there’s no scale, there’s no error or discussion of the data collection methods. But if you go to the site, the animation is mesmerizing. It went viral. Everyone loves it. In terms of generating interest and demonstrating big-picture concepts, this kind of medium can be very powerful.
This tip was one of the most interesting for me, as it gets into more of the process of doing science. Especially today, scientists generate a large amount of data. When I run the Community Earth System Model, I need more than a terabyte of data to input into the model. From there, I generate 2 TB of data that I use to input into a different configuration of the model from which I might get another 1 TB of data out. This is a huge amount of data to analyze, store, and make available for others to use.
As computing power increases exponentially and storage becomes easier and less expensive, petabytes of scientific data are getting stored. In many cases, it can be more efficient to buy additional storage space (like big hard drives) than to take the time to go back through old output and decide what to get rid of. But it’s important to think about how data is getting preserved during this process.
Data should be well documented with how it was collected or generated. Dates of processing and analysis should be included in the files and all contents should be labeled and explained clearly. A file with unexplained numbers is useless later. These are the common sense things that I’ve come to learn through the process of doing science.
The presenter also mentioned some other tips that I hadn’t thought of. A big one was to preserve data in plain text format. In climate science, we often use special file formats that give us access to analysis tools designed around those formats. But in order for others to make use of data for visualization, it’s important to provide a more globally-accessible format.
Finally, the speaker emphasized the need to document and share science tools. Her team goes a long way to communicate with scientists themselves in order to understand how they do their work. Without a design team, it’s still possible to communicate processes through the release of software and analysis procedures you use. This often happens in the publication of papers, but providing your code as open source is another way to achieve this.
The move to open source science has been gaining traction in the science community, in particular the need for transparency of analysis for the purposes of reproduction. Making code available is a viable way of doing this, but providing software to others can come with other caveats, including implicit support for others choosing to adopt that software for their own use. There is also the concern that if someone wants to reproduce a study, they should start with a blank slate instead of reusing parts of your tools which might be flawed in some way. I think transparency and accessibility is important in science, but both sides of the argument should be considered when deciding how each individual achieves this.
That was the long way around to get to the point that data visualization is important for science, but also not as easy as it may seem on the surface. There is a lot that goes into choosing the best way to represent data for public audiences. The SXSW Eco conference and this talk in particular demonstrated how designers can play a large role in effective communication in science. For a topic often misunderstood, like climate, they may play a more critical role than we think.