Text mining and real-time analytics of twitter data: A case study of australian hay fever prediction

Abstract

Social media platforms such as Twitter contain wealth of user-generated data and over time has become a virtual treasure trove of information for knowledge discovery with applications in healthcare, politics, social initiatives, to name a few. Despite the evident benefits of tweets exploration, there are numerous challenges associated with processing such data, given tweets specific characteristics. The study provides a brief of steps involved in manipulation Twitter data as well as offers the examples of the machine learning algorithms most commonly used in text analysis. It concludes with the case study on the Australian hay fever prediction with the application of the selected techniques described in the brief. It demonstrates an example of Twitter real-time analytics for heath condition surveillance with the use of interactive visualisations to assist knowledge discovery and findings dissemination. The results prove the potential of social media to play an important role in meaningful results extraction and guidance for decision makers.