Error in linear regression for multiple data-frames with matplotlib

Question

[Amr]

Error in linear regression for multiple data-frames with matplotlib

I want to make a 3 linear regression lines for different 3 data-frames one per each. there is one data-frame that has NA values maybe that's why an error has raised:

ValueError                                Traceback (most recent call last)
ValueError: On entry to DLASCL parameter number 4 had an illegal value

How to fix that error so that I am able to make 3 regression lines one per each data?

lo_sig = -1 *np.log2(df_significance['p.adjustMANOVA'])

lo_eff = -1 *np.log2(df_effect['p.adjustMANOVA'])

lo_sd = -1 *np.log2(df_sd['p.adjustMANOVA'])

df = pd.DataFrame({'Significance':lo_sig,'Effect_size':lo_eff,'Standard_deviation':lo_sd})

lo_sig2 = df_significance['s.dist']
lo_eff2 = df_effect['s.dist']
lo_sd2 = df_sd['s.dist']

df2 = pd.DataFrame({'Significance':lo_sig2,'Effect_size':lo_eff2,'Standard_deviation':lo_sd2})

ax.grid()


plt.scatter(y = df['Significance'],x = df2['Significance'], label='Significance', color='mediumseagreen',s=250)

plt.scatter(y = df['Effect_size'],x = df2['Effect_size'], label='Effect_size', color='indianred',s= 100)

plt.scatter(y = df['Standard_deviation'],x = df2['Standard_deviation'], label='Standard_deviation', color='orange',s = 20)

#Linear regression

a, b = np.polyfit(df2['Significance'], df['Significance'], 1)

plt.plot(df2['Significance'], a*df2['Significance'] + b,color = 'mediumseagreen', linewidth=4)

c, d = np.polyfit(df2['Effect_size'], df['Effect_size'], 1)

plt.plot(df2['Effect_size'], c*df2['Effect_size'] + d,color = 'indianred',linewidth=2)

e, f = np.polyfit(df2['Standard_deviation'], df['Standard_deviation'], 1)

#Here is the error

plt.plot(df2['Standard_deviation'], e*df2['Standard_deviation'] + f,color = 'blue',linewidth=4)

plt.show()

Thanks

Answer 1

[Amr]

The solution is to drop NA values and the regression line will be plotted, NA removed by df.dropna()