Evanston, USA, 22-24 June.
By Heather Packer, Agent and Research fellow, the University of Southampton.
The Web Science conference is really interesting because it brings together people from Computer Science, Artificial Intelligence, Web Engineering, Psychology, Economics, Law, Sociology, Ecology, Socio-cultural, and Media. The talks were really diverse, and included a lot of information about how people use various social media websites. The main message from the conference was that no matter how you design your social network, or social-network connected software, people will use it differently to how you expect, and that this may not be obvious until users habits are studied in detail. The focus of the research presented at the conference was how people used software, who is using it, and what they are doing.
The end June, I went to the ACM Web Science conference in Chicago, I was funded by the Software Sustainability Institute. If you are interested in applying for a fellowship which can fund you to travel from the SSI you can read more about it at: https://www.software.ac.uk/news/2012-07-11-funding-travel-and-more-new-fellowship-programme
Danah boyd (Microsoft Research) gave a great talk about security and privacy, and how people are developing their own practices that they find acceptable to maintain their personal data on the web. People are putting their data online, and instead of excluding files that are personal many people just give those files different access rights. Other people are more reserved by hiding messages in plain sight, with references to songs, films and exclusive codes between a group of friends. There was one case where a teenager only used Facebook in the evening and deactivated their account in the day, so that no adults could access it during the day. Other teenagers were using Twitter instead of Facebook to talk to their friends because it hasn't yet been adopted by their parents. The overall message was not to assume that the actual meaning of messages behind posts on social networks are known, and thus it may be harder to analyse how people's behaviour than it may seem .
Sinan Aral gave a keynote called "Content and Causality in Influence Networks", where he showed that products and applications can themselves be viral, rather than only the marketing around the product, through two aspects, "viral characteristics" and "viral features". Where the characteristics are things like usefulness and positive valence, and features such as invitations, notifications to friends and hypertext links . The talk was interesting because it showed how small changes in the ways that software is marketed can have massive differences in uptake and engagement levels - something that can really help when promoting your own software.
Clare Hooper (University of Newcastle), presented work on how HCI intersects with Web Science, and outlined four things we can learn from HCI that Web Scientists can learn from. The first, that like HCI, Web Science can be divided into two approaches, positivist and interpretivist. Positivist social scientists apply scientific methods to study society and interpretivist social scientists use social critique or symbolic interpretation rather than constructing empirically falsifiable theories. The second, research methods and methodologies need to be explained clearly, so that the assumptions of the underlying methods are correctly understood, in order to adapt them properly for other uses. The third, evaluation metrics are needed that are specific to Web Science, for example qualitative metrics to help us understand mobile computing and the emergence of Facebook and Twitter. Finally, design focus and design methods are required to exploit the interactions between millions of people .
Yahoo Research showed a search engine that shows results separated by whether the results are from left or right politically leaning sources . It therefore attempts to show what the search results in, depending on whether you want to see democratic party or republican party sources. For example, searching for "lies" shows that democratic sources talk about "glenn beck lies" and "fox news lies", whereas republican sources talk about "inconvenient truth lies" and "lies about obama". The talk highlighted how different news sources not only focus on different aspects, but also can even report completely different facts from a single news story. Their simple UI shows very quickly where bias lies on many issues, and is a great example of a simple user interface that delivers powerful information to the user.
Web Science is an emerging field with lots of emerging software and behaviour. This is highlighted by calls for approaches that support the design, evaluation and methodologies. It is also highlighted by how software is exploited to fulfil users' needs and what they decide is acceptable in terms of security and privacy. The overarching message of the conference was that no matter how you design your social network, or social-network connected software, people will use it differently to how you expect, and that this may not be obvious until users habits are studied in detail. It was refreshing to hear of such a significant amount of research being performed in the area of software, but focusing on how it is used, who is using it, and what they are doing.
1. Boyd, "Taken Out of Context: American Teen Sociality in Networked Publics", PhD Thesis (2008). http://www.danah.org/papers/TakenOutOfContext.pdf
2. Aral and Walker, Creating social contagion through viral product design: A randomized trial of peer influence in networks (2011) In: Management Science. http://icos.umich.edu/sites/icos6.cms.si.umich.edu/files/lectures/VPDFinal1110.pdf
3. Hooper and Dix, Web Science and Human-Computer Interaction: When Disciplines Collide (2012). http://www.clarehooper.net/files/papers/HCIWebSci.pdf
4. Yahoo! Labs, Political Insights Search: http://politicalinsights.sandbox.yahoo.com