ANALYZING TWITTER DATA OF FIRMS WITH SOCIAL MEDIA MINING

This study aims to determine whether Twitter data of the firms has a significant correspondence with respect to the firms, to cluster Twitter feeds of the firms and to find out which cluster has the maximum interaction through analyzing the Twitter data of the rival firms operating in different sectors. In this context, Twitter data shared by competitors operating in the cosmetics, electronics and marketplace sectors during 2017 were analyzed by following the process of Social Media Mining. The significant correspondence of Twitter variables of the firms was determined by the Correspondence Analysis. Twitter feeds of the firms were clustered with categories “Special Offer”, “Competition & Event”, “Product”, “Social”, “Support & Feedback” and “Special Interaction” by using a number of Text Mining pre-processing methods. Since the majority of the interactions obtained by the firms came from the minority of the feeds, which cluster received more interaction was analyzed with the help of the Pareto Principle