So we have very briefly looked at the background. We will revisit these themes throughout the blog but right now I want to explore the area of computational social science. How do we examine evidence differently and particularly how do we begin to explore more fully how the social sciences can exploit the masses of information we have available to us to help gather and analyze evidence differently.
I want to start with some of the work that influenced me in exploring this area. I should say that my interest is in socially and economically disadvantaged communities. I have been in public policy for a while and worked across government and the non profit sector. I became somewhat disillusioned with the culture of social science as a discipline especially the way in which evidence is collected and its application to practice. It seemed that while we had the ability to seek the right answers, the occupational culture that surrounds the collection and application of evidence is slow to keep up. And yet we are constantly confronted with ingenious examples of how to view social problems differently.
I became aware of the ability of new technologies to tell a real time story in Toronto. I was there in the summer of 2012 when there was a shooting at an outdoor party leaving one dead and a number wounded. BitchSlappedbylogic found the news and posted on Reddit. I was also reading a lot about what print and traditional media had to contend with when ordinary citizens could find the news first. It seemed interesting that social media provided a ready tool to find events and phenomena in a more efficient way. Quicker, more reliable, inexpensive, and directly from the people involved in those events themselves.
I also began delving more deeply into the ways in which science could support looking differently at evidence. Initially I was moved by an article I read in the Edge.org, it was a conversation with Nicholas A. Christakis about the potential of social science to learn from the physical sciences and vice versa. This debate has become a hot one over the last few years and a few related articles touch on the wider issues.
The article I read that got me on the road to thinking about these ideas more fully was by Nieman Journalism Lab contributor Jonathan Stray on the nature of algorithms, particularly its necessity in helping us shape how we perceive our world everyday. I was struck mostly by how necessary algorithms were in helping us shape our personalised content, the way we shop, read daily news, engage in our social networks, the list goes on.
So I began to think, if a person (Bitchslappedbylogic) could simply string together a list of events using hashtags and have such incredible impact, what could an algorithm do to better understand some of the complex social problems we face?
And then I came across another short interview with Alex (Sandy) Pentland of MIT about the power of Big Data, not to determine peoples beliefs but to ascertain people’s behaviour. You can imagine this was my lightbulb moment. Not because this isn’t obvious but because at the time not only had the shooting happened in Toronto, but then there was a more high profile shooting in Chicago involving a prominent hip hop artist. JoJo released a beef video aimed at Chief Keef and Lil Reese and was gunned down while riding his bike. Of course it all unfolded on Twitter, including confronting the subject of his ‘diss’ video twice in the street and filming it.
There had been longitudinal research done in Chicago, Peterborough (the UK) and Los Angeles done on the nature of their youth crime problem. All 3 were longitudinal studies on youth delinquency also asking the question why youth commit crime. Imagine an algorithm that could mine this publicly available data to ask new and different kinds of questions (the only one of the three in which the data is not publically available is the Peterborough study done by Cambridge). And further imagine that that data was combined with better understanding how an algorithm could design new kinds of information or evidence that could give us the answer to determining peoples behaviour in a way that could shed light on the issue. Like, if the main question was why people commit crime, and one of the glaring gaps in evidence is better understanding behaviour over time then surely social media and the data it provides would answer some of those questions.
The Guardian did interesting work during the London riots using Twitter as a tool of analysis. Again, it tells the same story from a completely different angle.
Hopefully I have given you some sense of the ways in which I am seeking to explore youth crime using openly available data in new ways. I hope to give some examples of what is already out there, though there isn’t lots there is enough to demonstrate my point.
I know, there are three strands here, the first is the one about what constitutes the evidence of social science, the second is what the police follow and record and the systematic nature in which they do it, and third is the way we engage the public in conversations about how they are affected and what we do about it.
- Defensive political science responds defensively to an attack on social science (andrewgelman.com)
- Seeing the World Through Soda Straws (greenconduct.com)
- Are the Social Sciences in Trouble? (washingtonmonthly.com)
- We need more mainstream social science, not less. (blogs.lse.ac.uk)
- Breaking down disciplinary boundaries by not building them (jonrainford.wordpress.com)