Project Overview

HOME

PROJECT OVERVIEW

PROJECT MANAGEMENT

ANALYSIS

FINAL DELIVERABLE

Company Information

Yelp is a ratings and recommendation website for businesses to connect with their users. Their main source of revenue comes from paid advertisements from businesses. To expand their business, they wanted to monetize the analytics that comes with the vast amounts data on Yelp. With that, they opened a new department for recommending solutions to businesses using analytics. Our group aims to make use of the given dataset to derive possible actionable solutions for businesses.

Problem Statement

At Yelp, we want our users to be matched with the best business, providing satisfied customers. Today, we are only able to promote the best matched business of choice to the users. However, this might not match the user’s expectations but only be the best of all existing choices. Therefore, we will use analysis in the area of text and sentiment to identify areas of improvement for businesses to be a better match to customer expectations. As customer expectations would be better met, we are hopeful that they would influence their friends to be a supporter of Yelp as well.

Primary Objective: Text and Sentiment Analysis

New Businesses

How can new businesses leverage on insights from the reviews and tips provided by the Yelp community to identify business opportunities? How can new businesses understand the trends, tastes and preferences of the country / state / city / neighbourhood to tailor their business propositions to best fit the preferences of users?

1. 1. What are the number of businesses under a unique category within a specific location? Is there a saturation of specific businesses? Is there a preference for specific operations within the area?
   2. Is there a trend for a specific type of restaurant within an area?
   3. Is there a lack of certain types of restaurants in the area?

Existing Businesses

Based on comparison of reviews and tips among competitors in the same business category, how can businesses find out the benchmarks and extract best practices to improve their businesses?

1. 1. What differentiates the popular businesses versus less popular businesses? What are the attributes consumers value (i.e. Happy Hour: True) that cause a business to become popular? What are the attributes that consumers value (i.e. Alcohol: None) which cause a business to become unpopular? (Generally, popular businesses have many check-ins and minimum business rating of 4 while unpopular businesses have few check-ins and max business rating of 2)
   2. What are the differences between the preferred attributes for different business categories / locations?

Secondary Objective: Data Visualisation

Interactive Dashboard for Sentiment Analysis Visualization

Visualisation of attributes / services that people look out for in businesses for specific country, state and city and business category based on sentiments.

1. 1. Word cloud
   2. Bar Charts for Rankings
   3. Pie Charts for Polarity Percentages
   4. Line Charts for time series analysis of ratings and change in preferences

Tertiary Objective: User Clustering

Prediction of Outlook for Business

How can Yelp leverage on the reviews, check-ins and tips provided by users to predict the success of the restaurants (ratings, positive reviews)?