Geo-Tagging Quality-of-Experience Self-Reporting on Twitter to Mobile Network Outage Events

Mobile wireless networks underpin digital economies and smart cities. Local and national scale network failures cause widespread social and economic impact. Self-reporting of consumer experience on social media platforms can inform operators. This paper investigates an innovative method to detect the consumer experience to outage events in both temporal and spatial domain using Twitter data. We use a variety of natural language processing (NLP) analysis to detect the consumer sentiment from a custom made dictionary and using naive Bayes classifier. We propose a hybrid geo-information extraction that sequentially extracts the geo-location from a priority list. A case study upon recent UK wide mobile network failure has been implemented in this paper. The results show that our proposed hybrid geo-information extraction system has been able to increase data size and accuracy of geo labelled Tweets. Also, our system can successfully detect this network issue in both time and location, which is validated by the national newspaper reports on this issue.