A retail company is training a BigQuery ML model to predict customer churn. The training data includes a `customer_id` column with high cardinality, a `signup_date` column, and a `total_purchases` column that ranges from 0 to 10,000. The data scientist wants to avoid overfitting and ensure the model treats the date as a meaningful trend rather than a categorical label. Which feature engineering approach should they use?

Drop `total_purchases`, keep `signup_date` as a string, and use feature selection to remove `customer_id`.

Drop `customer_id`, extract year and month from `signup_date`, and apply min-max scaling to `total_purchases`.

Label encode `customer_id`, convert `signup_date` to days since epoch, and apply log transformation to `total_purchases`.

One-hot encode `customer_id`, keep `signup_date` as a timestamp, and apply z-score normalization to `total_purchases`.

A financial services company has deployed a BigQuery ML model to predict loan default risk. The model was trained on data where the `income` feature was log-transformed and `age` was scaled to [0,1]. During batch prediction, the serving data contains raw `income` and `age` values. What should the data engineer do to ensure predictions are consistent with training?

Retrain the model on the raw serving data to avoid transformation mismatch.

Pass the raw serving data directly to the model; the model will handle the transformation internally.

Apply the same log transformation and scaling to the serving data before passing it to the model.

Use a different model that does not require feature scaling.

Google CloudProfessional Data Engineer

Domain 4Objective 2

4.2 Preparing Data for AI and ML PROFESSIONAL-DATA-ENGINEER Practice Questions (Page 5)

Part of the Preparing and using data for analysis domain, which accounts for ~15% of the PROFESSIONAL-DATA-ENGINEER exam. Google Cloud does not publish an official question count, but from its 120-minute exam (~50–80 total, ~8–12 in this domain), expect 3–4 from this objective — we provide 23 practice questions to prepare you well beyond it. (estimate)

23questions here

5free pages

5concepts

~15%of the exam

Questions 21–23

21application · medium
A retail company is training a BigQuery ML model to predict customer churn. The training data includes a `customer_id` column with high cardinality, a `signup_date` column, and a `total_purchases` column that ranges from 0 to 10,000. The data scientist wants to avoid overfitting and ensure the model treats the date as a meaningful trend rather than a categorical label. Which feature engineering approach should they use?
Select an answer first
22application · medium
A financial services company has deployed a BigQuery ML model to predict loan default risk. The model was trained on data where the `income` feature was log-transformed and `age` was scaled to [0,1]. During batch prediction, the serving data contains raw `income` and `age` values. What should the data engineer do to ensure predictions are consistent with training?
Select an answer first
23foundation · easy
In BigQuery ML, which feature engineering technique is used to transform a categorical column with high cardinality (e.g., product IDs) into a numerical representation suitable for a linear model?
Select an answer first

Finished these 3 questions?

Review the revealed explanations, or continue through the curriculum.

No more pagesBack to PROFESSIONAL-DATA-ENGINEER

Free Basic Practice is a study aid with revealable answers — not a scored exam. Examers.io is independent and not affiliated with or endorsed by Google Cloud. “PROFESSIONAL-DATA-ENGINEER” is a trademark of its owner, used for identification only.

4.2 Preparing Data for AI and ML PROFESSIONAL-DATA-ENGINEER Practice Questions (Page 5)

Questions 21–23

In BigQuery ML, which feature engineering technique is used to transform a categorical column with high cardinality (e.g., product IDs) into a numerical representation suitable for a linear model?

Keep studying — Preparing and using data for analysis