Abstract:

How to extract valuable insights from a noisy medium such as Twitter? Our approach is to use artificial intelligence to extract semantic features from the data. Once these features are extracted we can use machine learning techniques to extract valuable insights.?Our approach can be used to visualize how ideas are entangled inside a community's conversation as well as to identify the main themes in a corpus. Finally, it can be used to classify and track the evolution of specific topics in a Twitter stream. We'll provide guidelines and examples for the utilization of this methodology in a market-research context. Additionally, we introduce some applications of the proposed methodology to analyze two big 2016 social events: the U.S. presidential election and the Brexit referendum.