def scatter(df):

    plot_data(df, title="Data frame", yLabel="Time")

    plt.show()

    dr = compute_daily_return(df)

    plot_data(dr, title="Daily returns", yLabel="Daily returns")

    dr['GOOG'].hist(bins=20, label="GOOG")

    dr['SPY'].hist(bins=20, label="SPY")

    plt.legend(loc='upper right')

    # Scatterplot SPY vs GOOG

    dr.plot(kind='scatter', x = 'SPY', y = 'GOOG')

    spy = dr['SPY'][:-1] # remove nan value

    goog = dr['GOOG'][:-1] # remove nan value

    beta_goog, alpha_goog = np.polyfit(spy, goog, 1)

    # beta_goog= 1.23719057977

    # alpha_goog= -0.000283995818653

    plt.plot(dr['SPY'], beta_goog*dr['SPY']+alpha_goog, '-', color='r')

    plt.show()

    print("Correlation", dr.corr(method='pearson'))

    # Get kurtosis

    print("kurtosis=", dr.kurtosis())

if __name__ == '__main__':

    df=test_run()

    scatter(df[['SPY', 'GOOG']])