Amazee

Abstract

The team at FHNW has developed a Facebook crawler to integrate social data from users that log into Amazee via the "Facebook connect" single-sign-in facility. This way, Amazee can pre-populate the users profiles on Amazee without him/she having to enter that data manually. More importantly, it allows Amazee to collect relevant social data that can be used by Amazee's recommendation engine. This allows Amazee to make group and user recommendations right after a new user logs in, thus enhancing the probability that new users will stay on Amazee. During this effort on the crawler, the recommendation engine has been refactored and its performance enhanced, especially in regard to the computational speed; recommendations can now be made in real-time. This is especially important for recommendations to new users. The system has been thoroughly documented and handed over to the Amazee team; since a couple of months it is successfully in production. Another deliverable was the development of a social search functionality. The concept relies on social networks: if a user is logged into Amazee, her/his search results are ranked based on her/his social network position, meaning that pages or content from people that she/he knows ranks higher in the search results. This functionality relies on the reputation index that FHNW has developed previously. The feedback from Amazee has been very positive (Sehr gut, durch's Band besser).

In parallel, FHNW has developed together with Amazee a detailed specification and project plan for the implementation of a social cockpit (corresponding to work package 3 & 4, see Figure above). We have conducted several interviews with Amazee clients, designed wireframes and defined the information architecture and data definition. All features have been prioritized together with Amazee. Subsequently, the social cockpit has been successfully implemented and it was reviewed by Amazee. It is a sort of "Google Analytics" for social networks, customized for Amazee. It allows Amazee and its clients to get real-time statistics about their online communities, displayed in a very compact and efficient way. The social cockpit also allows to search for users and take actions (send email, block user, etc.). It therefore acts as a mass communication tool. The system has been developed as an extension of the current Amazee framework (Drupal), and therefore integrates smoothly in it.