How to plot linear regression with Seaborn based on a prediction of a target variable?

Question

I'm learning the very basics of data science and started with regression analysis. So I decided building a linear regression model to examine the linear relationship between two variables (chemical_1 and chemical_2) from this dataset.

I made chemical_1 the predictor (independent variable) and chemical_2 the target (dependent variable). Then used scipy.stats.linregress to calculate a regression line.

from scipy import stats

X = df['chemical_1']
Y = df['chemical_2']

slope, intercept, r_value, p_value, slope_std_error = stats.linregress(X,Y)
predict_y = slope * X + intercept

I figured out how to plot the regression line with matplotlib.

plt.plot(X, Y, 'o')
plt.plot(X, predict_y)
plt.show()

However I want to plot regression with Seaborn. The only option I have discovered for now is the following:

sns.set(color_codes=True)
sns.set(rc={'figure.figsize':(7, 7)})
sns.regplot(x=X, y=Y);

Is there a way to provide Seaborn with the regression line predict_y = slope * X + intercept in order to build a regression plot?

UPD: When using the following solution, proposed by RPyStats the Y-axis gets the chemical_1 name although it should be chemical_2.

fig, ax = plt.subplots()
sns.set(color_codes=True)
sns.set(rc={'figure.figsize':(8, 8)})
ax = sns.regplot(x=X, y=Y, line_kws={'label':'$y=%3.7s*x+%3.7s$'%(slope, intercept)});
ax.legend()
sns.regplot(x=X, y=Y, fit_reg=False, ax=ax);
sns.regplot(x=X, y=predict_y,scatter=False, ax=ax);

RPyStats · Accepted Answer · 2018-04-14 22:09:00Z

2

Using subplots and setting the axes will allow you to overlay your predicted Y values. Does this answer your question?

print(predict_y.name)
predict_y = predict_y.rename('chemical_2')
fig, ax = plt.subplots()
sns.set(color_codes=True)
sns.set(rc={'figure.figsize':(7, 7)})
sns.regplot(x=X, y=Y, fit_reg=False, ax=ax,scatter_kws={"color": "green"});
sns.regplot(x=X, y=predict_y,scatter=False, ax=ax, scatter_kws={"color": "green"});

edited Apr 14, 2018 at 22:09

answered Apr 14, 2018 at 18:45

RPyStats

3161 silver badge5 bronze badges

Sign up to request clarification or add additional context in comments.

4 Comments

samba Over a year ago

Updated my question, could you take a look at it? Also could you tell why the plot has different colors, not just blue as usual? Is it because of overlaying?

RPyStats Over a year ago

Your correct in that overlaying the plots will produce different colors, pass in additional arguments to sns.regplot can set them to be the same color. Y-axis name is getting overwritten with the second call to sns.regplot. It says 'chemical_1' as this is the name of the series predict_y. I can update my solution above.

samba Over a year ago

Got it, thanks! One more question: do you think I've added the legend correctly? It looks as expected but I wonder - is it the most optimal way or could it be improved?

RPyStats Over a year ago

Your last question is where it gets subjective, if your happy with it and it is working as expected then it's probably safe to keep it as is. Also, I haven't done much research on adding legends to seaborn plots.

Collectives™ on Stack Overflow

How to plot linear regression with Seaborn based on a prediction of a target variable?

1 Answer 1

4 Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

4 Comments

Your Answer

Sign up or log in

Post as a guest

Related