[Python] Doesn't imputing missing values effect the data quality?

I'm doing multiple linear regression for a dataset.

The numeric_df dataframe is the continuous variables in the orignal dataframe.

I want to check linearity between the price variable (target) and the other continuous variables through pairplots. When I draw pairplots after imputing missing values, the data is like this:



And when draw pairplots before imputing missing values, the data is like this:



Is the relationship still linear after imputing missing values? Wil a linear model work on it?

Continue reading...