Level: Intermediate
Mia Stephens, JMP Principal Product Manager, SAS
Ruth Hummel, JMP Senior Academic Ambassador, SAS
Are you wondering what you can learn from unstructured text data? Or are you just very interested in figuring out why people love Birkenstocks so passionately? In this talk we explore the world of text mining, using product reviews of Birkenstocks. First, we scrape the review data from the web, and then we use the Text Explorer platform to explore and analyze the data. We look at summary information and word clouds to understand the top terms used in the reviews, and use clustering to see terms that tend to co-occur and people who seem to say similar things. We use topic analysis to see if there are common themes in the reviews, and use data filtering to see if there are different themes for low and high-rated reviews. We then extract topics and other information from the text data, and use these structured data to develop statistical models that ultimately enable us to understand why people love their Birks so much. Finally, we compare reviews for Birkenstocks to reviews for the increasingly popular Crocs, to see if there are fundamental differences in consumer perception of the two brands.
