script 3.py

# Save the training data to CSV for the web app
training_data_csv = df.to_csv(index=False)

# Create a sample prediction function
def predict_dropout_risk(student_data):
    """
    Predict dropout risk for a student
    Returns: risk_level, probability, key_factors
    """
    # This would use the trained model in the actual web app
    # For now, we'll create a simplified version
    
    risk_score = 0
    key_factors = []
    
    # Attendance factor
    if student_data.get('attendance_rate', 100) < 70:
        risk_score += 3
        key_factors.append("Poor attendance (< 70%)")
    elif student_data.get('attendance_rate', 100) < 85:
        risk_score += 1
        key_factors.append("Below average attendance")
    
    # Academic performance
    avg_grade = (
        student_data.get('math_grade', 75) + 
        student_data.get('english_grade', 75) + 
        student_data.get('science_grade', 75)
    ) / 3
    
    if avg_grade < 60:
        risk_score += 3
        key_factors.append("Poor academic performance (< 60%)")
    elif avg_grade < 70:
        risk_score += 2
        key_factors.append("Below average grades")
    
    # Behavioral factors
    if student_data.get('disciplinary_actions', 0) > 3:
        risk_score += 2
        key_factors.append("Multiple disciplinary issues")
    
    # Family support
    if student_data.get('family_support', 5) < 4:
        risk_score += 2
        key_factors.append("Limited family support")
    
    # Homework completion
    if student_data.get('homework_completion', 80) < 60:
        risk_score += 2
        key_factors.append("Poor homework completion")
    
    # Determine risk level
    if risk_score >= 6:
        risk_level = "High Risk"
        probability = 0.8 + (risk_score - 6) * 0.02
    elif risk_score >= 3:
        risk_level = "Medium Risk"
        probability = 0.4 + (risk_score - 3) * 0.1
    else:
        risk_level = "Low Risk"
        probability = 0.05 + risk_score * 0.05
    
    return risk_level, min(probability, 0.95), key_factors

# Test the function
test_student = {
    'attendance_rate': 65,
    'math_grade': 55,
    'english_grade': 58,
    'science_grade': 52,
    'disciplinary_actions': 4,
    'family_support': 3,
    'homework_completion': 45
}

risk, prob, factors = predict_dropout_risk(test_student)
print(f"Test Student Prediction:")
print(f"Risk Level: {risk}")
print(f"Probability: {prob:.2f}")
print(f"Key Factors: {factors}")

print(f"\nDataset ready for web application!")
print(f"Total students in training data: {len(df)}")
print(f"Features used: {len(feature_names)}")