Cool Science: Crowdsourcing Big Data

Anyone who knows me at all knows I really like data.  It’s a tremendously nerdy interest, but I find data really fascinating, I guess in part because I love the idea that there is some great knowledge that’s hidden in the numbers, just waiting for someone to come along and dig it out.  What’s very cool is that we live in an age when technology allows us to generate massive amounts of data.  For example, the Large Hadron Collider generates more than 25 petabytes a year in data, which is more than 70 terabytes a day.  A DAY.  Some data analysis can be done by computers, but some of it really has to be done by people.  Plus, some studies really rely on the ability to gather data from massive groups of people in order to get an adequate sample from various groups to prove what you’re trying to show.  To solve these and other “big data” problems, some very smart and cool research groups have jumped on the crowdsourcing bandwagon and are having people from around the world get online and help solve the problems of data gathering and analysis.  Here are some cool projects I’ve heard about.

Eyewire: a group of researchers working on retinal connectomes at MIT found a fascinating way to get people to help with their data analysis – turn it into a game.  They have a good wiki that explains the project in depth, but the gist of it is that these researchers have microscopic scans of neurons from the retina.  Neurons are a huge tangled mess, so their computers could figure out how some of them fit together, but it really takes an actual person to go in and figure out what’s connected and what’s not.  So this team turned it into this 3D puzzle/game thing that’s really hard to explain unless you try it.  You go through a tutorial to learn how to use the system, and then you’re turned loose to start mapping neurons!  It’s not like the most compelling game I’ve ever played or something I’d spend hours doing, but it is interesting, and it helps neuroscience, so that’s pretty cool.

Small World of Words: this study aims to better understand human speech and how we subconsciously create networks of associations among words.  To do so, they set up a game to gather word associations from native and non-native English speakers.  Again, I wouldn’t necessarily call this a game in the sense of “woohoo, we’re having so much fun!” but it is kind of interesting to see what your brain comes up with when you’re given a set of random words.  (Plus it’s perhaps a little telling of your own psychological state if you really think about the words you’re coming up with.)  It takes like 2 minutes to do, and again, it’s contributing to science!  Also, according to their website, they are making their dataset publicly available, which as a research informationist/data librarian I wholeheartedly endorse.

Foldit: I haven’t played this yet, so I can’t speak to how fun it is (or boring), but it sounds similar to Eyewire in the sense of being a puzzle in which the players are helping to map a structure – in this case proteins.  Proteins are long chains of amino acids, but they fold up in certain ways that determine their function.  Knowing more about this folding structure makes it possible to create better drugs and understand the pathology of diseases.  For example, one of the things this project is looking at is proteins that are crucial for HIV to replicate itself within the human body.  Better understanding of the structure of these proteins could help contribute to drugs to treat HIV and AIDS.

So I encourage you to go play some games for science!  Do it now!  And if you’re at work and someone tries to stop you, just politely explain that you’re not playing a game – you’re curing AIDS.  🙂

Leave a Reply