Improving restaurants’ business performance using Yelp data sets through sentiment analysis

• Main Authors: Ching, Michelle Renee D., Bulos, Remedios De Dios
• Format: text
• Published: Animo Repository 2019
• Subjects: Sentiment analysis; Computational linguistics; Data mining; Restaurants; Computer Sciences
• Online Access: https://animorepository.dlsu.edu.ph/faculty_research/1697; https://animorepository.dlsu.edu.ph/context/faculty_research/article/2696/type/native/viewcontent
• Institution: De La Salle University

With the ever-present of social media sites and online review sites, access to customer’s sentiments and opinions on a business are within reach of any organization, which proves to be a gold mine of opportunity for them to provide the best customer experience. However, because of how enormous these data can be, such as the Yelp data sets that provides online reviews on different businesses, it will be difficult for them to extract valuable information, especially if they do not have expertise on doing so. Data analytics is a technique used to improve business productivity and gain through extracting, categorizing, and analyzing data to find meaningful patterns to provide the best experience for organization’s customers because it plays a significant role in motivating customer loyalty. Thus, this research study leveraged on sentiment analysis and opinion mining through the AYLIEN Text Analysis API that is available in RapidMiner data science tool, specifically the Aspect-Based Sentiment Analysis (ABSA) endpoint, performed a time series forecasting using linear regression for one year using Waikato Environment for Knowledge Analysis (Weka) machine learning workbench, and used the predicted data to conduct a linear regression in understanding the customers’ concerns of the five restaurants registered in Yelp to increase customer loyalty and profit through sustaining and/or improving customer satisfaction. Moreover, this research study recommends business strategies for the five restaurants based on the results of the one-year forecasted data using linear regression. © 2019 Association for Computing Machinery.