Order or StandardScaler and SimpleImputer

Module 1. The Predictive Modeling Pipeline
#1

What should be performed first on numeric data? StandardScaler() or SimpleImputer(missing_values=np.NaN, strategy=‘mean’)

I think inputting first will not affect the mean, but the standard deviation may be a bit lower. So the order should be 1. Standard Scaler, 2. Simple Imputer ?

#2

Yes, this is better to scale and then impute when possible. I gave some explanations in a early post where depending on the strategy it could be problematic: M1 Wrap-Up quiz Q5 SimpleImputer question - #3 by glemaitre58