Of Wines and Reviews: Measuring and Modeling the Vivino Wine Social Network
share
This paper presents an analysis of social experiences around wine consumption through the lens of Vivino, a social network for wine enthusiasts with over 26 million active users worldwide. We compare users' perceptions of various wine types and regional styles across both New and Old World wines, examining wines across price ranges, vintages, regions, varietals, and blends. We find that ratings provided by Vivino users are not biased by the cost of the wines. Then, we study how wine characteristics, language in wine reviews, and the distribution of wine ratings can be combined to develop prediction and recommendation models. More specifically, we model user behavior to develop three models: (1) a regression model for predicting wine ratings, (2) a classifier for determining user review preferences, and (3) a recommender system for suggesting wines based on history of ratings and reviews.
READ FULL TEXT