script.py

# Create synthetic training data for student dropout prediction
import pandas as pd
import numpy as np
from sklearn.model_selection import train_test_split
from sklearn.ensemble import RandomForestClassifier
from sklearn.linear_model import LogisticRegression
from sklearn.metrics import accuracy_score, classification_report, confusion_matrix
from sklearn.preprocessing import StandardScaler
import random

# Set random seed for reproducibility
np.random.seed(42)
random.seed(42)

# Generate synthetic student data
n_students = 1000

# Generate student features
data = {
    'student_id': range(1, n_students + 1),
    'age': np.random.randint(14, 19, n_students),
    'gender': np.random.choice(['M', 'F'], n_students),
    'socioeconomic_status': np.random.choice(['Low', 'Medium', 'High'], n_students, p=[0.3, 0.5, 0.2]),
    
    # Academic performance (grades 0-100)
    'math_grade': np.random.normal(75, 15, n_students),
    'english_grade': np.random.normal(78, 12, n_students),
    'science_grade': np.random.normal(76, 14, n_students),
    'previous_gpa': np.random.normal(3.0, 0.8, n_students),
    
    # Attendance (percentage)
    'attendance_rate': np.random.normal(85, 20, n_students),
    'days_absent': np.random.poisson(15, n_students),
    
    # Behavioral factors
    'disciplinary_actions': np.random.poisson(2, n_students),
    'participation_score': np.random.normal(7, 2, n_students),
    'homework_completion': np.random.normal(80, 25, n_students),
    
    # Family factors
    'parent_education': np.random.choice(['Elementary', 'High_School', 'College', 'Graduate'], n_students, p=[0.15, 0.4, 0.35, 0.1]),
    'family_support': np.random.randint(1, 11, n_students),  # Scale 1-10
    
    # School engagement
    'extracurricular_activities': np.random.randint(0, 4, n_students),
    'teacher_relationship': np.random.normal(7, 1.5, n_students),
}

# Create DataFrame
df = pd.DataFrame(data)

# Clip values to reasonable ranges
df['math_grade'] = np.clip(df['math_grade'], 0, 100)
df['english_grade'] = np.clip(df['english_grade'], 0, 100)
df['science_grade'] = np.clip(df['science_grade'], 0, 100)
df['previous_gpa'] = np.clip(df['previous_gpa'], 0, 4.0)
df['attendance_rate'] = np.clip(df['attendance_rate'], 0, 100)
df['participation_score'] = np.clip(df['participation_score'], 0, 10)
df['homework_completion'] = np.clip(df['homework_completion'], 0, 100)
df['teacher_relationship'] = np.clip(df['teacher_relationship'], 1, 10)

# Create dropout risk based on realistic factors
def calculate_dropout_risk(row):
    risk_score = 0
    
    # Academic performance factors
    avg_grade = (row['math_grade'] + row['english_grade'] + row['science_grade']) / 3
    if avg_grade < 60:
        risk_score += 3
    elif avg_grade < 70:
        risk_score += 2
    elif avg_grade < 80:
        risk_score += 1
    
    # Attendance factors
    if row['attendance_rate'] < 70:
        risk_score += 3
    elif row['attendance_rate'] < 80:
        risk_score += 2
    elif row['attendance_rate'] < 90:
        risk_score += 1
    
    # Behavioral factors
    if row['disciplinary_actions'] > 5:
        risk_score += 2
    elif row['disciplinary_actions'] > 2:
        risk_score += 1
    
    # Family and social factors
    if row['socioeconomic_status'] == 'Low':
        risk_score += 1
    if row['family_support'] < 4:
        risk_score += 2
    elif row['family_support'] < 6:
        risk_score += 1
    
    # Engagement factors
    if row['participation_score'] < 5:
        risk_score += 1
    if row['homework_completion'] < 60:
        risk_score += 2
    elif row['homework_completion'] < 80:
        risk_score += 1
    
    # Convert risk score to dropout probability
    if risk_score >= 8:
        return 'High Risk'
    elif risk_score >= 5:
        return 'Medium Risk'
    else:
        return 'Low Risk'

# Apply the dropout risk calculation
df['dropout_risk'] = df.apply(calculate_dropout_risk, axis=1)

# Create binary dropout column for machine learning
df['dropout'] = (df['dropout_risk'] == 'High Risk').astype(int)

print("Student Dropout Prediction Dataset Created")
print(f"Dataset shape: {df.shape}")
print("\nDropout Risk Distribution:")
print(df['dropout_risk'].value_counts())
print(f"\nActual dropout rate: {df['dropout'].mean()*100:.1f}%")

# Display first few rows
print("\nFirst 5 students:")
print(df.head())