We are all aware that our activity on the Internet can used without our consent for purposes that most of the users can’t even imagine. The website www.123people.com is a very well known website being able to find all the available information on the Internet for a person (the website searches and downloads all the photos, documents and websites on the Web and keeps records). Another example is Google. I remember when Google started providing its free DNS service competing with OpenDNS service. At that time, Google was accused by OpenDNS representatives that tries to extend the profiling of users it already does (records are kept by Google every time we search for a term using its search engine, all available search engines on the Internet do that – e.g. Yahoo, so they have to comply with the EU and US private data laws – for example they can only keep the records in some cases for a period of 6 months). The DNS service can offer an extended profiling by keeping records of every page the user visits. What about the social networks that we all use? What about the information that we share/upload on the Internet through social networks with our consent?
I was reading today the research which has been conducted from HP about social networks. A paper which was published recently, proposes some algorithms with which the preferences of the users on Twitter can be analyzed and finally forecast the box-office revenues for movies. The paper is titled “Predicting the future with social media”. The Internet is quite risky for social scientists as a means of conducting research since it limits the sample of the users, putting constraints that the researcher has to take into account in his/her analysis. Algorithms can be used on the other hand as a way of predicting simple outcomes (algorithms are used every time you visit websites trying to offer you the advertisements that interest you). Furthermore, relational behavior can be analyzed, since the users through social networks choose to interact only with specific users, not with all the users.