Measuring Cup is a little dataform project I've been working on this year. It's currently showing in Inside Out, an exhibition of rapid-prototyped miniatures at Object gallery, Sydney.
This form presents 150 years of Sydney temperature data in a little cup-shaped object about 6cm high. The data comes from the UK Met Office's HadCRUT subset, released earlier this year; for Sydney it contains monthly average temperatures back to 1859.
The structure of the form is pretty straightforward. Each horizontal layer of the form is a single year of data; these layers are stacked chronologically bottom to top - so 1859 is at the base, 2009 at the lip. The profile of each layer is basically a radial line graph of the monthly data for that year. Months are ordered clockwise around a full circle, and the data controls the radius of the form at each month. The result is a sort of squashed ovoid, with a flat spot where winter is (July, here in the South).
The data is smoothed using a moving average - each data point is the average of the past five years data for that month. I did this mainly for aesthetic reasons, because the raw year-to-year variations made the form angular and jittery. While I was reluctant to do anything to the raw values, moving average smoothing is often applied to this sort of data (though as always the devil is in the detail).
The punchline really only works when you hold it in your hand. The cup has a lip - like any good cup, it expands slightly towards the rim. It fits nicely in the hand. But this lip is, of course, the product of the warming trend of recent decades. So there's a moment of haptic tension there, between ergonomic (human centred) pleasure and the evidence of how our human-centredness is playing out for the planet as a whole.
The form was generated using Processing, exported to STL via superCAD, then cleaned up in Meshlab. The render above was done in Blender - it shows the shallow tick marks on the inside surface that mark out 25-year intervals. Overall the process was pretty similar to that for the Weather Bracelet. One interesting difference in this case is that consistently formatted global data is readily available, so it should be relatively easy to make a configurator that will let you print a Cup from your local data.
Sunday, June 06, 2010
Measuring Cup
Saturday, December 12, 2009
Data Walks - a #climatedata proposal
In response to the UK Met Office's recent data release and Manuel Lima's call for visualisations, there's been a flurry of #climatedata activity in the last couple of days, including some revealing visualisations. Though I'm looking forward to playing with the data myself, this isn't a post about visualisation. It's a simpler proposal for a way to make the data tangible.
Global warming is ultimately a question about change in a single measurement - temperature - over time. One way or another, it can be boiled down to a line graph. How best to make that line tangible? Visualisation is great, but how else could we feel those changes, especially over time? One way would be to walk the data. We could make a kind of giant line graph, in the form of a path or road, then walk from 1850 to 2009. According to the Met Office's graph - remixed above with a picture of my local landscape - this would be a fairly undulating journey, but the last half especially would be a distinct and noticeable climb. Building this path at a walkable scale seems like hard work though. It would be much easier to use the paths we already have. So, here's a recipe for a #climatedata walk:
- Make a graph. There are all kinds of options here. The Met Office graph shows global difference from a long-term (1961-1990) average. You could for example use local data only, or use raw average temperatures rather than difference from average. You would also need to select a year range from the data - want to walk the whole century or just post-WW2? All the data choices should be made clear to any walkers.
- Fit to landscape. This is the tricky part. The idea would be to find a walkable route with changes in elevation that fit your line graph well. Finding a perfect fit will be very difficult, but finding an OK fit should be possible. This will involve some scaling questions: how long will the walk be, and how much elevation will it cover? Accessibility, ergonomics, experience design, affect - lots of juicy design decisions here. One crude but easy fitting procedure would be to begin with a route, find its elevation profile, then scale the graph to fit the start and end points of the graph to the route start and end, then note the points where the path and the graph intersect. Maybe some GIS / maps people could help with software tools here for route finding and fitting?
- Tick marks. Walk the route and mark it out in order to make the whole thing legible. Mark out years or decades, as well as temperature variation (elevation). One option for paths with an imperfect fit, would be to notate the difference between the path and the graph at certain points, as well as points where the path and the graph intersect.
- Walk. Again you can imagine many ways to do this, ranging from big organised public walks, to smaller private ones. Of course walking often leads to talking - and in a different way to, say, looking at a graph.
