Airbnb: Data-Driven Stays

This project focuses on predicting Airbnb listing prices using machine learning techniques based on features such as location, property type, number of bedrooms, amenities, availability, and review scores.

The dataset underwent thorough preprocessing. Numerous outliers and missing values were handled through imputation—using the mean and median for numerical features, and the mode for categorical ones. Label Encoding was applied to transform categorical variables, and MinMaxScaler was used to scale numerical features to a standard range.

Exploratory Data Analysis (EDA) helped reveal trends and the most influential factors affecting listing prices. A set of regression models were trained and evaluated, including Linear Regression, Lasso, Ridge Regression, Decision Tree Regressor, and Random Forest Regressor. Model performance was assessed using metrics such as RMSE, MAE, and R² score.

The project successfully delivered a reliable predictive model, empowering Airbnb hosts and analysts with data-driven insights to set competitive and optimized listing prices. It highlights the real-world application of machine learning in the sharing economy and property rental sector.