Food and Beverage

Machine Learning - Michelin Star Restaurant Prediction

Model Accuracy: 90%

Tags: #Michelin #Logistic #Regression #Prediction #FnB #Asia #Restaurant #DataVisualisation


Since 1936, Michelin has written their guidebooks to award each restaurant they inspect with the range from 1 to 3 stars. The stars have since been recognized and respected as many chefs devoted their lives to pursue the award. With a single star, it can boost up to 20% more of restaurant business, 40% more business if the restaurant has two stars and 100 % more business with three stars.

However, the concealed inspection tends to give an uncertainty on how a restaurant is being inspected in order to avoid conflicts of interest. The prediction on how many stars a restaurant will get is highly favourable by investors and customers as the competition is getting tighter.


Analytico Asia used a dataset which includes the Michelin starred restaurants around the world. The dataset was then cleaned for the Asia region only. We used a machine learning model to predict which restaurant in Asia continent will receive a Michelin Star and achieved 90% of accuracy. The applied model can then be used to predict other continent as well as for further Exploratory Data Analysis (EDA).

Video Demo