Session: DAC-04-03: Data-Driven Design

Paper Number: 115206

115206 - Investigating Customer Preferences Using Video Product Reviews 

The wealth of online reviews has grown exponentially, attracting the attention of many researchers who recognize their potential as a valuable source of customer feedback. Leveraging online reviews enables product designers to gain a deeper understanding of customer preferences and make informed decisions to improve product design. Traditionally, the major source of online reviews comes from e-commerce websites, such as Amazon and eBay. However, as social media platforms gain popularity, video reviews from these platforms are becoming an increasingly important source of customer feedback. Video reviews have several advantages over traditional textual reviews. They are more comprehensive, less likely to be fake, have greater coverage of the customer base, and contain more interaction in the form of comments. The paper presents a four-stage methodology to analyze video reviews from social media platforms as an alternative source of customer feedback. This involves collecting and preprocessing video reviews, extracting product features using latent Dirichlet allocation (LDA) models, analyzing sentiment using the valence aware dictionary and sentiment reasoner (VADER) package, and computing feature importance using SHAP values. To the best knowledge of the authors, very few literatures have investigated the feasibility of using video reviews as a substitute for traditional online reviews. Therefore, this paper contributes to the growing literature by demonstrating the viability of video reviews as an alternative source of online reviews to understand customer preferences and their implications for engineering design.

Presenting Author: Kangcheng Lin University of Illinois

Presenting Author Biography: Kangcheng Lin is a graduate student.

Authors:

Kangcheng Lin University of Illinois
Harrison Kim University Of Illinois