New methods for knowledge discovery in geo-localized social media networks

Resumen

In this thesis, we have addressed several problems pursuing a common objective: developing methods to discover knowledge based on the SM activity located in an area during a period of time and the information available about the authors of this SM activity (be it derived from the Social Network they are part of or inferred from their behavior towards other users). More specifically, three main objectives have been pursued in the present thesis: - The design, analysis and implementation of new methods to quantify the impact of different topics and events for a particular location over time exploiting the analysis of geo-localized SM interactions. - The creation and implementation of new approaches to support the near-real time monitoring of campaigns and the quality of service for a location, based on the insights inferred from the analysis of SM interactions attached to this particular location. - The design and implementation of knowledge discovery systems that deliver operationalizable metrics to support industry use cases based on our methods. The first two objectives have been profoundly addressed in 4 publications (2 per objective), being the third one orthogonal to all publications: Impact quantification methods on locations: - A new model to quantify the impact of a topic in a location over time with Social Media - Quantifying the emotional impact of events on locations with Social Media Geo-localized Campaigning and Quality of Service Monitoring: - CARESOME: A System to Enrich Marketing Customers Acquisition and Retention Campaigns Using Social Media Information - Leveraging Localized Social Media Insights for Industry Early Warning Systems The results we obtained in the tests we performed after implementing the required knowledge discovery systems, proved the performance of our methods to support different industry relevant use cases, especially in the Marketing, Sociology, Politics, etc domains.