Yelp Business Analysis and Location Recommendation

Preview Link(Deployed URL): https://cmpt732-demo-web.azurewebsites.net

Demo Video: https://www.youtube.com/watch?v=j9IToR9_agg

Menu

• Yelp Business Analysis and Location Recommendation
• Menu
  • Introduction
  • Run the application locally
  • Repo Structure
  • Tech Stack

Introduction

The Yelp challenge dataset was chosen because we want to work on a scope with potentially real business needs. After doing a simple explanatory analysis of the dataset and integrating the feedback from the project proposal, our team decided to initiate a project on offering data analysis targeting business owners and potential investors by concentrating on the following products: a dashboard for an overview of the market, an analysis of users’ reviews using natural language processing to provide insight, and a machine learning algorithm offering recommendation about where to choose the location for the chain restaurants. A specific region, Alberta province was chosen for project demonstration.

Run the application locally

See the RUNNING.md for more details.

Repo Structure

• flask-backend: Serves as the backend for our web visualization page.
• frontend: Contains the react repo for our frontend webpage.
• neighbors: Contains the machine learning logic for the location recommender part.
• nlp: Contains the NLP part to analyze the user review for each business.
• restaurant: Collect and clean the business data.

Tech stack

• Data Processing: Spark, Spark SQL, Pandas.
• Machine Learning: Scikit-learn.
• Database: MongoDB.
• Web backend: flask .
• Web frontend: React.js, Chart.js, leaflet.js, Leaflet.markercluster.js.