Multifaceted Visualisation of Annotated Social Media Data

Social media (such as online communities) is one of the main sources generating large, unstructured, and redundant big data. Analytics of social media data helps to maximise its utility, and visualisation plays a significant role in the exploration of both big data and the results of the analysis on it. A large amount of big data from online communities (and other social networks) are typically visualised through their interactions (e.g., who interacts with whom) in the form of graphs where people are represented as nodes and the interactions between them as edges. But there is a lot of other information that could be analysed and visualised, such as annotations performed through crowd sourcing which has been very popular in recent time. How can one visualise the social network data from the prism of such annotations? This is what we address here. Spring embedding algorithms have been used in semantic visualisations depicting semantic similarity relationship among nodes. In this paper, we borrow this idea and propose a multifaceted visualisation of annotated online community data using spring embedded graphs. The unique feature of our approach is that it offers a visualisation that captures the proximity of the members to annotated concepts in a heterogeneous community graph. We have evaluated our approach to visualise annotations related to emotions and work barriers faced by members of a Government run online community that was trialled to provide informational and emotional support to its members. We show with specific examples how our approach offers a multifaceted visualisation of the online community and facilitates interpretations and analysis of the large amount of social data captured.