correlation

corr.py

def correlation(df, v1, v2):
    """
    Calculates the correlation between two
    variables in a pandas dataframe.
    """
    # Get mean of each column
    mean1, mean2 = df[v1].mean(), df[v2].mean()
    
    # Get squared deviations from mean for each column
    squared_deviation1 = [(x - mean1)**2 for x in df[v1]]
    squared_deviation2 = [(x - mean2)**2 for x in df[v2]]
    
    # Get the variance of each column
    var1 = sum(squared_deviation1) / (len(df) - 1)
    var2 = sum(squared_deviation2) / (len(df) - 1)
    
    # Get the standard deviation of each column
    std1, std2 = np.sqrt(var1), np.sqrt(var2)
    
    # Calculate Z scores of all values, i.e.
    # distance from mean in terms of standard deviations
    z1 = ((df[v1] - mean1) / std1).copy()
    z2 = ((df[v2] - mean2) / std2).copy()
    
    # Calculate the correlation coefficient!
    return sum(z1 * z2) / (len(df) - 1)