Linear Regression statsmodel: Missing a required outcome variable

Question

I am using statsmodel to build a linear regression for salary. It's giving me an error that says I am missing a required outcome variable. Can you tell what I am doing wrong here? My initial model, my_model seemed working but the ols fitted model says I am missing a required outcome variable.

my_model = str('Salary ~  X1 + X2')`
my_model = str('')

train_model_fit = smf.ols(my_model, data = X_train).fit()

print(train_model_fit.summary())
X_train['predict_salary'] = train_model_fit.fittedvalues
X_test['predict_salary'] = train_model_fit.predict(X_test)

Answer 1

[Answering for completeness]

You set the model formula in the first line:

my_model = str('Salary ~  X1 + X2')`

But then you assign an empty string to it:

my_model = str('')

Hence the error.