Archive for the ‘Data Analysis’ Category

Data Collection Isn’t Boring

November 16, 2011 Leave a comment

This title isn’t a straw man: I don’t mean by it the obvious claim that experimental methods are interesting. With all the current talk about the deluge of data we face, though, I am thinking about the adventure, creativity, and big science power that goes into data collection. Maybe we feel like we have too much data now–but we also have new methodologies and the promise of extracting quantitative data even from fields like literature and the arts. We start with the data in hand in most statistics classes and ask how we can analyze it: but behind bland terms like “missing data” and “spatial statistics” lie amazing stories about fieldwork, the invention of new experimental methods and the martialing of forces to organize huge data collection projects.

The first time this was brought home to me, I was sitting in the Harvard Statistics lounge waiting for a section to start and leafing through old issues of Chance, the magazine about statistics aimed at laymen. It’s a monument to how statistics really impacts our world, with articles on current policy and scientific issues, computing, and  how statistics are presented. The most fun part is the human interest stories: what caught my eye that day was an article about the dangers of data collection in the tropics among poisonous creatures, and how to get the best data possible given the limits of tropical fieldwork.

Recently, a networks exhibit at the Hall of Science and my reading have given me a few more cool examples of data collecting tricks to think on. Walter Tschinkel is an entomologist at Florida State who has a unique way of studying ant colonies and their social structure: pouring molten metal into the colonies and waiting til it hardens to see the shape of the tunnels (more here). The resulting molds make for great data and great art–much like the cloud chamber tracks that grace another area of the museum exhibit hall. In my reading of Mayr, he mentions some other tactics for studying elusive populations: the most vivid of which is spraying pesticide into the forest canopy and doing a biodiversity census by classifying the bugs and birds that rain down from the trees.

Right now we have international data collection efforts that are changing the face of science: particle physics at CERN, genome projects, efforts towards a comprehensive Encyclopedia of Life. We’ve realized that our challenge now is to make data meaningful and usable, and we’re rising to that challenge. It reminds me of some of the very first international science efforts, like the British Navy’s collaboration with stations throughout the world to record the tides under Whewell’s direction (after he convinced everyone of how useful detailed tide data would be), and the days when logarithm and integral tables enabled all kinds of new science. Behind so many advances in science are new ways of looking at data that have an aesthetic interest all their own.

Categories: Data Analysis, Evolution