Towards a More User-Centric Perspective in Utilising 'Big Data'

The next speaker at "Compromised Data" this afternoon is Asta Zelenkauskaite, who notes the increasing interweaving of social and mainstream media; based on the properties of 'big data' it therefore becomes important to explore how users engage with mass media and cross-media contexts. How relevant are 'big data' to the mass communication field?

Traditional media outlets have been mainly focussing on a quasi-passive engagement with media content, while social media now offer a two-way interaction by providing back channel functionality. Mass media content, user-generated content, and user interactions' digital imprints are coming together to shape this cross-media environment.

The first aspect of 'big data' is the sheer size of the data on user activities, of course; a second is the varity of activities, and a third the velocity of engagement. Veracity is also increasingly being highlighted, and in the end it is especially the value of 'big data' which needs to be explored. Questions here are first, value to whom?, and second, value of what?

Value may lie for example in interactivity and interest-based content discovery. Interactivity is increased through 'big data' environments, and this also enhances information discovery processes. Asta uses the example of Italian radio in this context: Italian radio stations emphasise the various social networks which can be used to engage with the radio stations and to discover additional media content - yet also notes that interaction with such tools is quite marginal, provides only limited content discovery opportunities, and increases source fragmentation.

This raises questions for possible content architecture: the current model for radio stations and mass media in general is one of top-down content access, where the mass media outlet controls the mainstream channel and user-generated channels are used only marginally and exist in the form of various, fragmented, competing spaces (which are also difficult to monetise because of this fragmentation).

An alternative model would provide for interest-based content access which brings together various information items of varying provenance in the same curated navigation space, enabling users to engage with a broader variety of content on their own terms. This requires us to conceptualise content in a fluid way, building on an interest-based information architecture matrix which users can navigate freely regardless of the differences between media outlets included in this matrix.

The value of 'big data' in this context would lie in more interest-based content discovery, increased interactivity, and more content variety across multiple media content streams. For researchers, it would also enable more cross-media analysis that builds on these data. But there is also the problem of potential information overload for users taking advantage of such cross-media systems; of ethical issues with the data veracity in this environment; and of a greater potential for data-enabled surveillance of users.

This is an argument for thinking about the value of 'big data' from a more user-centric perspective, then. Value extraction through user-centric approaches is still in its infancy, and the proprietary nature of the data makes this process even more difficult