simulate_attract_personality.py

def simulate(simulations=1000, n=500, quantile=0.80):
    
    correlations = np.zeros(simulations)

    for i in range(simulations):
        
        data = {"personality": np.random.randn(
            n), "attract": np.random.randn(n)}
        df = pd.DataFrame(data)
        
        df['personality_attract'] = df['personality'] + df['attract']
        cutoff = df['personality_attract'].quantile(quantile)
        
        df['outcome'] = np.where(
            df['personality_attract'] > cutoff, 'dated', 'rejected')
        
        dated = df.query('outcome == "dated"')
        
        correlations[i] = dated.corr()['personality'][1]
        
    return correlations