Big Data, Compromised Data?

The final panel at AoIR 2015 for today is the Compromised Data panel, celebrating the release of the book of the same name. Ganaele Langlois starts us off by highlighting the key themes of the book: data are now crucial to building the social, and the gaps and omissions in the data therefore have very significant impacts.

Governments such as the Harper government in Canada, which has just been removed from office, seek to hide some forms of data because it is politically expedient to do so, and so data are political; data are also shaped by the sources we draw them from, however – such as the social media platforms we work with.

We cannot therefore uncritically embrace 'big data', but need to reflect on the conditions and consequences of our work: data are now being mined for profit, and this often squeezes out publicly-funded, public-value research. And our data are social; we cannot necessarily expect them to make any sense without taking into consideration the social life of data – indeed, our own analysis of the data contributes to building the social.

Large-scale datasets are largely approached using data visualisation, of course, and such visualisations are far from unproblematic; they introduce distortions and selections. Data could be understood as immersive sensoriums, and we might thus think about how we curate reality by representing the data in a variety of ways.