Modelling the Yelp Rating of Restaurants: Disparity of Rating Between Different Restaurant Cuisines and Prices

Overview

This repo contains code and data used in the writing of the paper.

We use the Yelp dataset to explore and model how different attributes of a restaurant play a factor in the review ratings. We found there are some variables which are promising predictors for the rating of a restaurant. Cuisines such as Chinese or Mexican tended to have lower ratings, and lower priced cuisines performed worse.

File Structure

The repo is structured as:

• data/analysis_data contains the cleaned dataset that was constructed from Yelp Dataset.
• models contains fitted models.
• other contains the map we've created from the dataset, details about LLM chat interactions, and sketches.
• paper contains the files used to generate the paper, including the Quarto document and reference bibliography file, as well as the PDF of the paper.
• scripts contains the R scripts used to simulate, download and clean data. As well as creating the interactive map.

Statement on LLM usage

LLM such as ChatGPT was used to assist writing scripts. Further details on the usage is documented in other/llm/usage.txt. No other LLM usage was made.

Downloading dataset

The original dataset was too big to include in this repo. Visit https://www.yelp.com/dataset (Last accessed: 2024 April 15), then click "Download Dataset", fill out the personal information, and download the JSON (NOT the photos). Extract and place the files in data/raw_data folder (there should be PLACE_EXTRACTED_YELP_DATASET_HERE file there).

Interactive Map

Interactive map for the raw data can be access here: https://alexsohn1126.github.io/yelp-analysis/other/map/rawmap.html.

And the map for the cleaned data can be accessed here: https://alexsohn1126.github.io/yelp-analysis/other/map/map.html. Try clicking on datapoints, it shows which restaurant it is, and its rating, cuisine, and price range!