Predicting Business Viability of Restaurants

To survive in the restaurant business, getting ahead of the competition becomes mandatory. With the help of data analysis tools, managing the information becomes a simple task. One stands to sharpen its competitive edge, increase profitability, boost profit margins, and grow a customer base with the right data, and the right software. But with tremendous demand, there always arises the risk of saturation.

The demographic data, behavioral data, and the interests shared by customers will help the business in forecasting trends, and provide a complete view of the business venture so that innovative strategies to improve operational efficiency, customer satisfaction, and business customer base develop profitability. The data would be very useful for any new restaurants to predict future customer preferences and tastes to pioneer strategics techniques and successfully use it for competitive advantage.

In this use-case, we aim at identifying the factors that help the successful operation of a restaurant, by leveraging models in Machine Learning and Data Science techniques to identify the type of cuisine served, type of the restaurant, location of restaurant, facility of online ordering is available or not, etc.

We used machine learning models and data science techniques to gain insights that help a restauranter in determining the factors in building a successful restaurant in the Bengaluru location, India.

We use regression models of machine learning models to understand which model has predicted with utmost accuracy.

• 1. Linear regression
• 2. Lasso regression
• 3. Ridge regression
• 4. Random forest

We have observed that Random Forest is yielding us the best accuracy score with 89%. And We also observed that ridge and Lasso have very low scores. After model fitting, we found out the predicted rating of a restaurant actually matches in 91% cases in real-time.