Getting Started with Machine Learning in Nepal: A Complete 2025 Guide

# ============================================================
# Complete ML Pipeline with scikit-learn
# Predicting house prices using linear regression
# ============================================================

import numpy as np
import pandas as pd
from sklearn.datasets import fetch_california_housing
from sklearn.model_selection import train_test_split
from sklearn.preprocessing import StandardScaler
from sklearn.linear_model import LinearRegression
from sklearn.ensemble import RandomForestRegressor
from sklearn.metrics import mean_squared_error, mean_absolute_error, r2_score
import warnings
warnings.filterwarnings('ignore')

# ──────────────────────────────────────────────────────────
# STEP 1: Load Data
# ──────────────────────────────────────────────────────────
print("=" * 50)
print("STEP 1: Loading Data")
print("=" * 50)

data = fetch_california_housing(as_frame=True)
X = data.data      # Features: MedInc, HouseAge, AveRooms, etc.
y = data.target    # Target: median house value (in $100k)

print(f"Dataset shape: {D}{"{X.shape}"}")
print(f"Features: {D}{"{list(X.columns)}"}")
print(f"Target range: {D}{"{y.min()}":.2f}k - {D}{"{y.max()}":.2f}k")
print()

# ──────────────────────────────────────────────────────────
# STEP 2: Exploratory Data Analysis
# ──────────────────────────────────────────────────────────
print("STEP 2: Basic EDA")
print("-" * 40)
print(X.describe().round(2))
print()

# Check for missing values
print(f"Missing values: {D}{"{X.isnull().sum().sum()}"}")
print()

# ──────────────────────────────────────────────────────────
# STEP 3: Train / Validation / Test Split
# ──────────────────────────────────────────────────────────
X_trainval, X_test, y_trainval, y_test = train_test_split(
    X, y, test_size=0.2, random_state=42
)
X_train, X_val, y_train, y_val = train_test_split(
    X_trainval, y_trainval, test_size=0.25, random_state=42
)

print(f"Train size:      {D}{"{X_train.shape[0]}"} samples")
print(f"Validation size: {D}{"{X_val.shape[0]}"} samples")
print(f"Test size:       {D}{"{X_test.shape[0]}"} samples")
print()

# ──────────────────────────────────────────────────────────
# STEP 4: Feature Scaling
# ──────────────────────────────────────────────────────────
scaler = StandardScaler()
X_train_scaled = scaler.fit_transform(X_train)
X_val_scaled   = scaler.transform(X_val)
X_test_scaled  = scaler.transform(X_test)

# ──────────────────────────────────────────────────────────
# STEP 5: Train Models
# ──────────────────────────────────────────────────────────
lr_model = LinearRegression()
lr_model.fit(X_train_scaled, y_train)
lr_val_pred = lr_model.predict(X_val_scaled)

rf_model = RandomForestRegressor(
    n_estimators=100, max_depth=10, random_state=42, n_jobs=-1
)
rf_model.fit(X_train, y_train)
rf_val_pred = rf_model.predict(X_val)

# ──────────────────────────────────────────────────────────
# STEP 6: Evaluate on Validation Set
# ──────────────────────────────────────────────────────────
def evaluate(name, y_true, y_pred):
    mse  = mean_squared_error(y_true, y_pred)
    rmse = np.sqrt(mse)
    mae  = mean_absolute_error(y_true, y_pred)
    r2   = r2_score(y_true, y_pred)
    print(f"  RMSE: {D}{"{rmse:.4f}"} | MAE: {D}{"{mae:.4f}"} | R2: {D}{"{r2:.4f}"}")

print("Validation Performance:")
evaluate("Linear Regression", y_val, lr_val_pred)
evaluate("Random Forest     ", y_val, rf_val_pred)
print()

# ──────────────────────────────────────────────────────────
# STEP 7: Final Evaluation on Test Set
# ──────────────────────────────────────────────────────────
rf_test_pred = rf_model.predict(X_test)
evaluate("Random Forest", y_test, rf_test_pred)

# ──────────────────────────────────────────────────────────
# STEP 8: Feature Importance
# ──────────────────────────────────────────────────────────
importances = pd.Series(
    rf_model.feature_importances_,
    index=X.columns
).sort_values(ascending=False)

for feat, score in importances.items():
    bar = "X" * int(score * 50)
    print(f"  {D}{"{feat:15s}"} {D}{"{bar}"} {D}{"{score:.4f}"}")