TAGGING MENU ITEMS FROM USER REVIEWS
Intent:
- Two Searches not supported by Food based platforms
> Best dishes of a restaurant as per customers
> Best restaurants serving a particular dish
Proposed Solution:
- Get food mentions from user reviews for restaurants over Zomato
- Associate these mentions with menus of respective restaurants
How to Run:
1.Navigate to Zomato-Food-Review folder in terminal
2.Run localhost server using 'php -S localhost:8000'
3.On browser goto 'localhost:8000/UI/foodie.html', enjoy the platform :)
Data:
- Zomato API calls for dailymenu doesn't work, as restaurant menus as images
- Thus, we manually took menus of few restaurants as image to text was inefficient
- Also there is a limit of fetching reviews using Zomato API
- We thus, scraped reviews from zomato site, though we have a Zomato API wrapper
:Codebase Description-
zomato-api-wrapper/:
1.zomatoApiModule.py - a python wrapper to the zomato API
2.driver.py - driver file for zomatoApiModule.py classes
backend/:
1.menu_generator.py - Creates restaurant_id to name and menu mappings from menu.csv
2.readfile.py - Combine Review and Ratings for Individual restaurants
3.Match.py - Algorithms to match mentions with menu items
4.POStagger.py [Rule based NER] - Create food mentions from user reviews
4.Client.py - Creates Search by restaurant and Search by dishes mappings for UI.
5.data/ - Intermediate data of backend
6.Ratings/- Ratings by user reviews for restaurants
7.Reviews/ - Reviews of restaurants
8.menu.csv - menu of some restuarants as our dataset
visualize/ [Create Visualization Inferences from data]:
1.visualize_1 - Relate Average Restaurant Price per dish, Words per review length
and Restaurant Rating
2.visualize_2 - Relate Price, Popularity and Ratings of individual dishes