Abstract:
The ubiquity of social media, blogs and various internet news sources allow users to collaborate, share thoughts, express concerns and vent anger on matters that distress them. Gathering and analysing data from various social media platforms provide governments with the opportunity to understand issues that citizens are most concerned about in society and discover issues of interest. It is challenging mainly due to the complex nature of social media data. In this thesis, the main research problem is the design of an SMA framework for governance - that is to mine and make sense of social media data to support decisions in governance. The framework is implemented and evaluated using guidelines provided by Design Science Research (DSR) methodology. Well-proven decision support systems and recommender systems frameworks are used for the foundation of the system, to analyse social media data and provide recommendations to users. In this multi-methodological research, implementation of the prototype enables the conceptual framework to be evaluated by discovering any problems with the model. Evaluating the implemented system was done using self-evaluation, using real-world data feeds from Twitter and analysing that data using the system to discover breaking news and situations successfully. The theoretical contribution of this research consists of the requirements and the conceptual framework. The practical contributions of this research include the system architecture and the prototype used for analysis of social media data. The main limitation of this research was that the implementation is an evolving prototype and must be further developed for use. Hence, only one data source (Twitter) was used when developing the system. It would be beneficial to use additional data sources (e.g. news articles and Facebook group posts) to enhance the system and to corroborate data from other sources. Furthermore, the algorithms used were not scientifically evaluated to discover the best-performing ones before using them in the prototype, although they were selected for their proven appropriateness for the task. Future research could be directed to improve the capability of the prototype, integrate more data source, and to support multiple languages.