Real Estate Price Prediction Pipeline

Data Acquisition: Loads the dataset from CSV, standardizes column names to lowercase, and displays initial data distributions and shapes.

Data Cleaning & Wrangling:Maps and normalizes column names (e.g., bedrooms, bathrooms, size).Uses regular expressions (Regex) to extract numeric values from text fields (price, size, rooms).

Filters records specific to New Cairo and drops rows missing essential metrics (price, size_sqm).

Removes pricing outliers by filtering based on the 10th and 90th percentiles of price-per-square-meter.

Feature Engineering & Preprocessing:Creates a custom feature (room_size_factor).Handles categorical variables using One-Hot Encoding (pd.get_dummies).

Splits data into training and testing sets, imputes missing values using the median, and standardizes features using StandardScaler.

Model Building & Training: Trains an XGBoost Regressor with fine-tuned hyperparameters (n_estimators=1000, learning_rate=0.05, max_depth=6).

Overfitting Analysis: Generates learning curves using 5-fold cross-validation to monitor training vs. validation performance.Model Evaluation: Computes the $R^2$ score and Mean Absolute Error (MAE), applying a conditional scaler if needed to present real-world EGP values clearly.Diagnostic Visualizations: Plots three critical charts side-by-side using Seaborn/Matplotlib:Actual vs. Predicted values.Feature Importance (top price drivers).Error Distribution (Residuals).