Implicit Social Production: Utilising Socially Generated Data By-Products

Enhancing business processes by the integration of social software is an area of active research. Once such integration has occurred, a new problem is presented - that of using social data in an effective manner. With large amounts of user generated data created, finding relevance in both data and in the people who created it as part of a business process becomes problematic. This paper frames the problem of socially generated information in the context of Open Source software development processes and of improved execution of tasks in that domain. Such social processes highlight the research area of facilitating the automatic selection of relevant data as part of a larger process. The paper introduces a novel two stage mechanism to answer such a problem. The approach is built on the concept of using the implicit social connections available from socially generated data artefacts to create a weighting model. This methodology is inherently egalitarian in nature as it uses a folksonomical strategy to construct the model. A dynamic domain specific lexicon is created to improve term weighting relevance. This weighting is then enhanced by analysing implicit proximity between participants of the socially generated production. By combining these two methods within a software framework, finding relevancy within a large corpus of socially generated data is improved. The prototype software framework built on these two approaches is constructed to provide dynamic programatic access to social data which can be incorporated as part of a larger business process to speed up the decision making process.