“Lovely evening today in Chicago isn’t it Madam! So I’ve come to U Chicago to learn to do computation social science. Now what on earth does that mean?
Well, how did you get into town this morning? You tapped into public transport perhaps? Well, that is a piece of data. And I see you have a smartphone in your hand Madam? Perhaps you just sent out a tweet about your favourite deep dish restaurant? That’s another data point. And maybe you downloaded a song to listen to on those headphones? More data. Now times that by the 20 odd million people in this city and you’ve got a lot of information. If I could analyse even a fraction of that, I could learn a lot of about Chicago life. That’s the social science bit.
In order to do that, I need to be able to “speak” to a computer to tell it how to analyse this data. That’s the computational bit.
It’s very exciting!”
I was quite happy with my impromptu bit of whimsical role play. The interview panel… well, I was placed on the wait list. I guess they just didn’t believe the weather could be lovely in Chicago.
However, three months later, I was bumped up and awarded a Fulbright scholarship, for which needless to say I feel impossibly blessed, and six months later I’m sitting in a sub-letted room in South Chicago, on the morning of my first class as part of U Chicago’s brand new MA in Computational Social Science (MACSS), contemplating the fact that I’m about to go Back to School…
By instinct, I value my privacy (if I was that Chicagoan, I’d have walked away like “why does this stalker wanna know how I got into town?"), but I’ve decided to share some thoughts on my journey on this program somewhat publicly.
a) save time
b) solve problems
c) create fun content
2). Use this skill to help make understanding society
a) more simple
b) more beautiful
Hence, I hope I can be an example to others who may not consider themselves particularly tech-savvy or tech evangelists, but may still want to learn computer programming to achieve practical goals!
Me: “How is the music project going?”
I wasn’t quite prepared for her response:
“We couldn’t figure out how to get pairwise differences down to two dimensions (it was one or like 16), we’ve moved on from the idea of a genre map to metrics involving single-dimension distances. Right now, we’re working on some MPI code to distribute the pairwise-distance calculations to several nodes in an effort to parallelize the job. Even if we use just a 1% sample of the database, we’ll be computing nearly 50 million pairwise distances. We’ve already written a bunch of mapreduce scripts that will take the pairwise distances and output certain things (like the two most similar songs in the dataset, for example).”
After expressing my complete bemusement, she assured me:
“Guaranteed you will know what all those terms mean at this time next year.”
I look forward to returning to this post in June 2018 to see if her confidence is warranted, and if so whether I can explain it in lay terms! If so, I will be quietly satisfied with my first year on the program.