438 Miles Up: Analyzing Urban Nature from Orbit
By Alex Kindel
B.S. Symbolic Systems (Learning), 2014
Read about the CityNature project on the OutWest student blog. Over the summer, a team of undergraduate student researchers combined spatial analysis with innovative mining of planning document text, photographs, social media, and published historical narratives to explain why nature is unevenly distributed in and across cities.
How is nature distributed in cities? In what ways can we understand the quality and experience of urban nature? These are just some of the many questions tackled by the interdisciplinary City Nature project this summer, which I've had the great fortune to be a part of.
To answer these questions, we've taken a variety of approaches, from historical explorations of city parks to data mining on city planning documents. For my portion of the project, I chose to take a quantitative approach, primarily using methods from remote sensing. My main data source is Landsat 5, a satellite orbiting 438 miles above Earth. Landsat 5 carries an instrument called the Thematic Mapper (TM), which captures and processes numerous wavelengths of light reflected from the planet's surface. Using this imagery, I've spent the summer exploring and quantifying greenness for 36 of the largest cities in the United States.
My typical day in the office starts early in the morning in the computer lab, where I do the kinds of analysis that require a lot of heavy lifting on the computer's part. Landsat imagery is corrected for satellite conditions, transformed to generate greenness measures, split into smaller zones, and analyzed for statistical error. After spending a few hours in the lab, I take a lunch break, then head back to the shared office where the summer research assistants usually work. In the afternoon, I take the measures I've generated and explore different ways of categorizing and visualizing them.
To me, one of the coolest parts of this project has been seeing how ubiquitous computational methods are in the kind of research I'm doing. I regularly work with datasets containing up to 200,000,000 values at a time, so when I want to do any sort of quantitative analysis, automated systems are required to make it happen. Scientific computing isn't just good for raw power, however; I can also use automated methods to cluster my data into sensible categories, or to visualize data so I can more easily draw my own qualitative conclusions from it. This technical aspect to my research is something I find particularly fascinating, especially as a Symbolic Systems major. In fact, I probably spend as much time studying the theory and math behind my methods as I do applying them to my data!
My summer is winding down now, but it's definitely been a rewarding experience. Seeing applications for the computational techniques I study in my classes has only made my fascination grow, and I'm excited to see what more can be done when scientific computing and humanistic questioning collide!
Read more at the Out West Blog for Summer Interns »
