What are Web Observatories?

“Do you know what a web observatory is?” demanded our lecturer. “Er  – a place where you watch what’s happening on the Web?” volunteered a classmate.  Basically, although we’d heard of ‘Web observatory methods’ while preparing for our dissertations, it became clear that few of us had thought through what these might entail beyond a vague idea of sitting around checking out other people’s Facebook activity. In fact, it’s a ‘global data resource for the advancement of economic and social prosperity.’

In the words of that same lecturer, “To keep pace with the Web’s growing scale and scope, Web Science research demands the development of new theories, the availability and interpretation of relevant data, effective and scalable multilevel analytical methods, and considerable computational infrastructure.”  So the Observatory is an online, mixed-methods, interdisciplinary environment for collaboration and sharing focusing on data about the Web. It provides tools and methodologies to examine data and activity. And to make it more complicated, it’s also a web of observatories, with 15 different Observatories co-ordinated by the Web Science Trust at Southampton.

The type of questions that concern Web Observatories range from inward-looking interrogation, such as what is the taxonomy of a Web Observatory, to outward-facing investigation, such as how can the Web itself be used as a tool to study ‘real world’ events.

One such web methodology is ‘Living Analytics’, based on the analysis of real time interaction between people online. One use of this looks at collaborative filtering – a way of creating personalised recommendations – and suggests that this outperforms matrix factorisation techniques on several dimensions, including accuracy. A little bit more ambitious than sitting around on Facebook, then.