Pfthb

I'm trying to plot a scatter plot of the values of actual sales (y) and predicted sales (ŷ).

I have imported the csv file and currently the codes I have for the linear regression model is:

result = smf.ols('sales ~ discount + holiday + product', data=data).fit()
print(result.summary())

Since, I only have the actual sales values, how do I find the predicted sales (ŷ) values to plot the scatter plot? I have tried researching and found lm.predict() and result.predict(). Is there a difference? lm = LinearRegression()
Thank you in advance!

Without data it is hard to help, but I guess you have X and y from dataset because you want to perform linear regression. You can split data into training and test set using scikit-learn:

from sklearn.cross_validation import train_test_split
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size = 1/3)

Then you need to fit linear regression to the training set:

from sklearn.linear_model import LinearRegression
regressor = LinearRegression()
regressor.fit(X_train, y_train)

and afterwards predict test set results:

y_pred = regressor.predict(X_test)

Finally, you can plot your test or training results:

# Visualising the Training set results
plt.scatter(X_train, y_train, color = 'red')
plt.plot(X_train, regressor.predict(X_train), color = 'blue')
plt.title('Discount vs Sales (Training set)')
plt.xlabel('Discount percentage')
plt.ylabel('Sales')
plt.show()

# Visualising the Test set results
plt.scatter(X_test, y_test, color = 'red')
plt.plot(X_train, regressor.predict(X_train), color = 'blue')
plt.title('Discount vs Sales (Test set)')
plt.xlabel('Discount percentage')
plt.ylabel('Sales')
plt.show()

(In this scenario we want to predict how many Sales will be if we set specific value of e.g. Discount percentage). If you have more than one X parameter, things are more complicated and you will need to use dummy variables, perform statistical analysis etc..