Efficient Feature Transformation | Dataford Interview Questions

Given a dataset, write a function to perform efficient feature transformation.

Represent the dataset as a list of row dictionaries and standardize each selected numeric feature using z-score normalization. Return a new dataset, preserve unselected fields, and replace constant-feature values with 0.0.

Input and Output

dataset is a non-empty list of dictionaries, and columns contains numeric feature names present in every row. Return a new list of dictionaries with transformed values.

Constraints

1 <= len(dataset) <= 10^4
1 <= len(columns) <= 20
Every selected feature exists in every row
Feature values are finite numbers
The original dataset must not be modified

Function Signature

def transform_features(dataset, columns):

Your question is Efficient Feature Transformation. Start with the requirements in the Question tab.

Run and submit as often as you like. When you're ready, talk me through your approach or go straight to the code.

Problem