This work package will integrate and advance the methods for mining social big data, within the SoBigData research e-Infrastructure. The expected outcomes are methods that are simultaneously principled (carry strong theoretical guarantees) and practically applicable, providing a broad statistical foundation for Big Data Analytics. The results of WP9 will endow and progressively enrich the RI, and thus the offerings of WP6 TA and WP7 VA, with interoperable scalable methods and services for analysing large-scale social data, as needed to produce comprehensive analyses of complex socio-economic phenomena.
The specific objectives are:
• To enhance the repertoire of efficient multi-dimensional analytical methods and algorithms for social mining.
• To integrate, systematize and make accessible analytical processes to understand social phenomena (e.g., personal behavioural profiles, the collective behavioural patterns, global patterns of diffusion and social influence, patterns of how sentiments and opinions vary in our societies).
T9.1. Social Network Analysis. SoBigData develops a framework of multi-dimensional/multi-layer network analytics, which integrates and enhances the plurality of methods of the national infrastructures (see WP6 for details).
T9.2. Social Pattern Discovery. This task enhances the SoBigData e-Infrastructure with new methods and services for discovery of emergent patterns and rules, describing socio-economic behaviour of interesting sub-populations.
T9.3 Text mining and semantic annotation of big social data. This task enhances the SoBigData e-infrastructure and the underlying GATE TSMM infrastructure, by integrating a wide range of methods for clustering, mining and semantic analysis of unstructured social content, especially of short length such as Facebook posts, tweets, blogs, and even search-engine queries.
T9.4 Mobility data analytics. This task advances and integrates in SoBigData a rich repertoire of methods and services for the analysis of human mobility, based on the world-leading expertise of CNR, UNIPI, FRH and AALTO in mobility data mining.
T9.5 Visual analytics. This task advances and integrates in SoBigData a suite of scalable visual analytics methods for interactive visually-driven context-aware analysis of mobility and social network data, as well as visual exploration and interpretation of the results from the text analytics and social pattern discovery methods and services.
T9.6 Social behaviour and human dynamics. This task enhances the SoBigData infrastructure with methods and services for measuring and understanding social behaviour and human dynamics.
T9.7 Engineering Massive Scale Social Mining. This task addresses the major computational challenges of big data analytics and social mining, such as the high volume and dimensionality of data, the dynamics and velocity of information generation and the requirements of real-time reactions to changes in the environment.