df = pd.read_csv('data_heart_disease.csv')
# df = pd.read_csv('https://raw.githubusercontent.com/ash322ash422/data/refs/heads/main/data_heart_disease.csv')

df

print(f"Dataset shape: {df.shape}")

df.head()

target_counts = df['target'].value_counts()
target_counts

# Lets visualize

fig, axes = plt.subplots(1, 2, figsize=(10, 4))

# Target balance
target_counts = df['target'].value_counts()
axes[0].bar(['No Disease (0)', 'Disease (1)'], target_counts.values,
            color=['steelblue', 'tomato'], edgecolor='white', linewidth=1.5)
axes[0].set_title('Target Class Distribution', fontweight='bold')
axes[0].set_ylabel('Count')
for i, v in enumerate(target_counts.values):
    axes[0].text(i, v + 1, str(v), ha='center', fontweight='bold')

# Age distribution by target
df.groupby('target')['age'].plot(kind='kde', ax=axes[1], legend=True)
axes[1].set_title('Age Distribution by Outcome', fontweight='bold')
axes[1].set_xlabel('Age')
axes[1].legend(['No Disease', 'Disease'])

plt.tight_layout()
plt.show()

df

X = df.drop(columns=['target'])   # everything except the label
y = df['target']                   # the label

print(f"Features  X: {X.shape}")
print(f"Labels    y: {y.shape}")
print(f"\nFeature names: {list(X.columns)}")

X
# y 

corr = df.corr()
corr

# Heatmap allows you to read above numbers visually
fig, ax = plt.subplots(figsize=(10, 7))
corr = df.corr()
mask = np.triu(np.ones_like(corr, dtype=bool))   # show only lower triangle

sns.heatmap(corr, mask=mask, annot=True, fmt='.2f', cmap='coolwarm',
            center=0, linewidths=0.5, ax=ax, annot_kws={"size": 8})
ax.set_title('Feature Correlation Matrix', fontweight='bold', fontsize=13)
plt.tight_layout()
plt.show()

# Highlight features most correlated with target
print("\nTop features correlated with 'target':")
print(corr['target'].drop('target').abs().sort_values(ascending=False).to_string())

from sklearn.model_selection import train_test_split

X_train, X_test, y_train, y_test = train_test_split(
    X, y, test_size=0.2, random_state=42
)

print(f"Training samples : {X_train.shape[0]}")
print(f"Test samples     : {X_test.shape[0]}")

# 1) Which columns have missing values?

missing = X_train.isnull().sum()
print("Missing values in train per column:\n", missing)


missing = X_test.isnull().sum()
print("\nMissing values in test per column:\n", missing)

missing = X_train.isnull().sum()
print("Missing values in train per column:\n", missing)


missing = X_test.isnull().sum()
print("\nMissing values in test per column:\n", missing)

from sklearn.ensemble import RandomForestClassifier

model = RandomForestClassifier(n_estimators=100, random_state=42)
model.fit(X_train, y_train)

print("Model trained successfully!")
print(f"   Trees in forest : {model.n_estimators}")
print(f"   Features used   : {model.n_features_in_}")

# Lets test on a sample:
print(list(X_test.iloc[0]))
print(y_test.iloc[0])

sample = [[57.0, 1.0, 0.0, 150.0, 276.0, 0.0, 0.0, 112.0, 1.0, 0.6, 1.0, 1.0, 1.0]]

y_pred = model.predict(sample)
y_pred

from sklearn.metrics import accuracy_score, classification_report

y_pred = model.predict(X_test)

accuracy = accuracy_score(y_test, y_pred)
print(f"Test Accuracy: {accuracy:.2%}\n")

print("Full Classification Report:")
print(classification_report(y_test, y_pred, target_names=['No Disease', 'Disease']))

print("actual :", list(y_test)[:10])
print("predict:", list(y_pred)[:10])

from sklearn.metrics import confusion_matrix, ConfusionMatrixDisplay

cm = confusion_matrix(y_test, y_pred)
print("confusion matrix:\n", cm)

fig, axes = plt.subplots(1, 2, figsize=(11, 4))

# Raw counts
disp = ConfusionMatrixDisplay(cm, display_labels=['No Disease', 'Disease'])
disp.plot(ax=axes[0], cmap='Blues', colorbar=False)
axes[0].set_title('Confusion Matrix (Counts)', fontweight='bold')

# Normalised (percentages)
cm_norm = cm.astype(float) / cm.sum()
disp2 = ConfusionMatrixDisplay(np.round(cm_norm * 100, 1), display_labels=['No Disease', 'Disease'])
disp2.plot(ax=axes[1], cmap='Blues', colorbar=False)
axes[1].set_title('Confusion Matrix (%)', fontweight='bold')

plt.tight_layout()
plt.show()

importances = pd.Series(model.feature_importances_, index=X.columns).sort_values(ascending=True)

fig, ax = plt.subplots(figsize=(8, 5))
colors = ['tomato' if v == importances.max() else 'steelblue' for v in importances.values]
importances.plot(kind='barh', ax=ax, color=colors, edgecolor='white')
ax.set_title('Feature Importances (Random Forest)', fontweight='bold')
ax.set_xlabel('Importance Score')
ax.axvline(importances.mean(), color='gray', linestyle='--', alpha=0.7, label='Mean importance')
ax.legend()
plt.tight_layout()
plt.show()

print("\nTop 5 most important features:")
print(importances.sort_values(ascending=False).head())

# Took 
from sklearn.model_selection import GridSearchCV

# Took 20 + minute
# param_grid = {
#     'n_estimators'    : [100, 200, 300],
#     'max_depth'       : [None, 5, 10, 20],
#     'min_samples_split': [2, 5, 10],
#     'min_samples_leaf' : [1, 2, 4],
#     'max_features'    : ['sqrt', 'log2']
# }

# Took 1 minute
param_grid = {
    'n_estimators'    : [100, 200],
    'max_depth'       : [5, 10, ],
    'min_samples_split': [2, 3,],
    'min_samples_leaf' : [1, 2],
    'max_features'    : ['sqrt',]
}



grid_search = GridSearchCV(
    estimator  = RandomForestClassifier(random_state=42),
    param_grid = param_grid,
    cv         = 5,            # 5-fold cross-validation
    scoring    = 'accuracy',
    n_jobs     = -1,           # use all CPU cores
    verbose    = 1
)

grid_search.fit(X_train, y_train)

best_model = grid_search.best_estimator_
best_acc   = best_model.score(X_test, y_test)

print("Best Hyperparameters:")
for k, v in grid_search.best_params_.items():
    print(f"  {k:25s}: {v}")

print(f"\nBaseline accuracy  : {accuracy:.2%}")
print(f"Tuned model accuracy: {best_acc:.2%}")
print(f"Improvement        : +{(best_acc - accuracy)*100:.1f} pp")