Data, Metadata, Big Data?

I'm spending the day at the Centre for Communication and Computing at the University of Copenhagen, where Klaus Bruhn Jensen has brought together a bunch of AoIR folks, including myself, for a one-day symposium called "Digital Data - Lost, Found and Made". I'll be speaking about our Twitter research in the afternoon.

Klaus begins by noting that there are three degrees of media: humans as media, the mass media, and network media - and our understanding of media and their flows has changed considerably over the years. From the two-step flow model we've now moved to the three steps of one-to-one, one-to-many, and many-to-many communication processes. Each of these have a prototype: face-to-face, print and broadcast, and online communication (especially following Web 2.0 models), although examples which differ considerably from these prototypes also exist.

Communication generates data, and most of these data are being - have been - lost. Some can be found again, through what is essentially a form of media archaeology: texts are the legacy of humanity. But data can also be made, through observation and interaction: this is the legacy of the social sciences. In the case of media and communication, this also includes metadata, and has a history which includes, but also stretches well beyond, recent approaches to the study of audiences from TV meters through to data mining. These are precursors to the current buzz about 'big data'.

The computer itself is a meta-medium: it emulates and remediates past media, thereby also affecting these media, and it provides the basis for the development of entirely new media forms - it supports both the adaptation of old media into new media, as well as the development of 'born digital' new media.

Finally, there is meta-communication, which establishes the conditions under which we communicate - such as the definitions of the terms we use, and the relations between the actors who are communicating. Metadata bear witness to these forms of meta-communication; they make explicit these processes.

One tool for research in this field is social network analysis: it positions networks as globally measurable structures which determine the distribution of data and metadata. From an optimistic perspective, everything is documented in such quantitative analysis - but what is lost or missed with such global approaches are often the local contexts and specificities which may only be examined through further qualitative work.

This is a key challenge for the research approaches now emerging in this area, as 'big data' becomes the new research paradigm.