A Simple Method to Impute Missing Ratings

Suppose we are given the movie ratings data given by viewers, represented in a matrix $(2-$D numpy array$).$ Admissible ratings are integers in the range $1-10.$

However, most users have not rated most movies, so the entries corresponding to those missing entries are zero, which actually mean that such ratings are unknown $($not given$).$ In this exercise, we will predict those ratings by simply averaging the columns and imputing the missing values.