RSS feed Get our RSS feed

News by Topic

BizReport : Research archives : December 18, 2014

Researchers use Yelp to predict a restaurant's chances of survival

Researchers at the University of Maryland have developed a tool that appears to show a connection between the success of a restaurant and the content of Yelp reviews.

by Helen Leggatt

By analyzing more than 2,000 Washington D.C. restaurants open as of December, 2013, including 450 that had closed down between 2005 and 2014, along with a database of more than 130,000 Yelp reviews, UMd researchers claim to be able to predict a restaurant's chances of survival with 70% accuracy.

Using software that can "read" the text content of reviews, totaling 15 million words or more in this research, the team were able to identify linguistic patterns that foretold closure by comparing restaurants based on price, cuisine and descriptions.

"The star ratings are definitely important, but to look at the text as well brings a more nuanced ranking," said assistant professor Shawn Mankad. "The text is bringing new information to the table."

The researchers focused on words associated with restaurants that have been able to remain in business. They found restaurants with reviews that included words such as "food," "good," "place, "like," "order," "friend," "time," "great," "nice," and "service" usually survived at higher rates.

The working paper, 'More Than Just Words: Using Latent Semantic Analysis in Online Reviews to Explain Restaurant Closures', grew out of a dissertation by UMd doctoral student Jorge Mejia.

"The whole idea is that we are surrounded by all of this free, unstructured data," said Mankad - hundreds of thousands of words that would require armies of employees to read, let alone interpret. "We should be using that data."

Image via Shutterstock

Tags: online reviews, research, restaurant

Subscribe to BizReport



Copyright © 1999- BizReport. All rights reserved.
Republication or redistribution of BizReport content is expressly prohibited without the prior written consent.
BizReport shall not be liable for any errors in the content, or for any actions taken in reliance thereon.