Incorporating Online User-feedback into Design of the iPhone: From Sentiment Analysis to Feature Definition

Open Access
Author:
Cornell, Brennan Frederick
Area of Honors:
Engineering Science
Degree:
Bachelor of Science
Document Type:
Thesis
Thesis Supervisors:
  • Gul Kremer, Thesis Supervisor
  • Melik C Demirel, Honors Advisor
  • Judith A Todd, Faculty Reader
Keywords:
  • Multiple-Generation Products
  • Web-Based Review
  • Emotion
  • Sentiment Analysis
  • Text Mining
  • Forum Mining
Abstract:
Changing consumer opinion is a barrier to Multiple Generation Products (MGPs). MGPs are products that keep core functionalities through iterations but experience incremental improvement with each iteration. This is a common strategy that allows companies to maximize profits, but failure to accommodate consumer’s needs can still be a detriment to a successful MGP. Limited research suggests systematic methods of incorporating user-feedback into the design process. Consumer reviews and online forums represent a wealth of untapped consumer opinion data. The research presented in this thesis defines a systematic method of gathering and processing consumer feedback from the web and makes recommendations for integrating the data into the design process. This thesis discusses, in detail, the author’s previous conference work (1), where 2515 online reviews were collected and processed by the algorithm proposed. The algorithm maps consumer sentiment to a visualization that may be incorporated into the design process for understanding of design needs. Implementation of the visualization would aid in the design process by process quick, reliable feedback to designers.