Booking cancellations have substantial impact on hotel management systems. It affects the services, decrease revenue, and limits the production of accurate booking forecasts. To overcome the effect of cancellations, some hotels use strict cancellations policies and define overbooking tactics. However, these regulations have a negative impact on number of bookings and on hotel reputation which in turn leads to revenue loss. The goal of this project is to minimize the losses by using the power of data science and machine learning algorithms to build a classification model to predict which bookings are likely to cancel. This is helpful for revenue management, supplies purchases and pricing decisions as well.

I used the Hotel booking demand dataset available in https://www.kaggle.com/jessemostipak/hotel-booking-demand. It contains more than 120,000 observations and 32 features. The data are retrieved from two different hotels: Resort hotel and City hotel. A few feature highlights include :

• lead_time: number of days that elapsed between the entering date of the booking into the PMS and the arrival date
• previous_cancellations: number of previous bookings that were cancelled by the customer prior to the current booking
• deposit_type: indication on if the customer made a deposit to guarantee the booking this variable assumes three categories: No Deposit / Non Refund / Refundable.

• NumPy • Pandas • Scikit-learn • Matplotlib • Seaborn