Harvesting Consumer Opinion and Wine Knowledge Off the Social Media Grape Vine Utilizing Artificial Neural Networks

In this paper we mine over 80 million twitter microblogs in order to explore whether data from the social media initiative known as Twitter can be used to identify sentiment about red wines. We test to see whether models derived from Twitter data can corroborate industry sales figures and we employ text analysis software developed to assess emotional, cognitive, and structural components of text to analyze the twitter dataset to harvest knowledge about consumer sentiment on different wine varietals. A multi-knowledge based approach is proposed using, Self-Organizing Maps and domain expertise in order to establish view the social network conversation. We show that it is possible to both confirm previously known knowledge and find novel information through the proposed methodology.