Analyzing product attributes and brand sentiment of smartwatches using Twitter/X data from a time series perspective

This research paper delves into Twitter data analysis through hashtag searches associated with smartwatches, offering a framework for extracting product attributes and sentiments from a time series perspective. A sample of 133,000 tweets was collected from Twitter in two distinct periods (t1 and t2) to scrutinize the prevailing sentiments and product attributes evident in online chats about smartwatches. This study aims to uncover valuable insights into brand sentiment and product attributes by comparatively analyzing brand sentiment and word clouds, as well as employing Latent Dirichlet Allocation (LDA) to identify topic evolution between time periods t1 and t2. The outcomes of this investigation highlight the significance of employing text analytics as a potent method for gauging consumers' opinions concerning emerging product attributes from a time series perspective. The study also provides procedures and actionable recommendations for businesses, elucidating how they can harness text data to gain a deeper understanding of consumer perceptions pertaining to their products and those of their competitors from a time series perspective.