Measuring interactivity and geographical closeness of Online Social Network users to support social recommendation systems

Several applications (e.g., Instagram, PiCsMu) integrate existing Online Social Networks (OSN) into the core of their solutions to explore social information. Although this integration enables more accurate social recommendation systems, the collection and monitoring of relevant OSN data by third-party applications is a challenging management task, since OSNs (a) impose rate restrictions to their Application Programming Interface (API) calls, (b) do not provide detailed information about specific OSN features, and (c) may provide incomplete or not up-to-date OSN data. Therefore, this paper covers the design, prototyping, and evaluation of JSocialLib, a new meta-API library for collecting OSN data from existing OSNs. It provides (1) an interaction- and (2) a location-based method in support of social recommendations systems.