AirBnB Price Analysis

This project aims to explore and predict the price of Airbnb listings based on various attributes. With over 7.7 million listings across more than 100,000 cities worldwide, Airbnb is a key player in the global short-term and vacation rental market, generating $9.9 billion in revenue in 2023. The platform allows property owners to list their homes, while users browse and book accommodations based on their preferences.

While Airbnb simplifies property rentals for hosts and users, determining the optimal price for listings remains a challenge. Hosts must decide on a price that maximizes their profits while considering various property features and local market conditions. On the other hand, users seek better insights into pricing variations to make informed decisions.

In this project, we analyze data from the InsideAirbnb project, which includes detailed information on over 100,000 listings across five European cities. Our goal is to provide valuable insights on price predictions by examining factors such as property features, location, and local demand. We start by preprocessing the data for consistency, then build predictive models using explainable features. We evaluate the accuracy of these models and compare the predicted prices to actual prices, helping both hosts and users make more informed decisions.

Contents:

• Data
  • Overview of Data
  • Full Preprocessed Data
  • Data Scraping and Preprocessing
  • Feature Engineering
• Models
  • Overview of Models
  • Explainability with SHAP
  • XGBoost Regressor
  • Feature Importance with SHAP
• Results
  • Model Performance
  • XGBoost Regressor Performance and Feature Importance
  • Finnaly the fainess map of the listings in London