Question 1

A company wants to process customer purchase data in near real-time to update their analytics dashboard. The data should go through some transformation and enrichment steps before being loaded into BigQuery. Which of the following pipeline designs would best meet this requirement?

Options :

A : Use Pub/Sub to capture streaming data from the application, then use Dataflow to transform and enrich the data before loading it into BigQuery.

B : Write a Cloud Function to read data from a file in Cloud Storage whenever the file is updated, process it, and load it into BigQuery.

C : Set up Cloud Dataprep to transform the data, then use a batch job to load the data into BigQuery every hour.

D : Set up a Cloud Scheduler job to trigger the pipeline every minute, pulling data from the database, transforming it, and loading it into BigQuery.

Answer: A

Question 2

A marketing analytics team wants to track the effectiveness of their digital campaigns using Looker. They need a dashboard with a line chart displaying the click-through rate (CTR) trend over time, segmented by campaign, and want the ability to filter by date range. Which steps would best allow you to implement this dashboard while meeting the requirement for interactive date filtering and campaign segmentation?

Options :

A : Create a pivot table with CTR as a row and date as columns, and filter campaigns within the pivot table.

B : Create a line chart with date on the x-axis, CTR on the y-axis, and set campaign as a series to segment the data. Add a date filter to the dashboard.

C : Create a custom visualization in Looker that directly links to campaign sources for additional drill-down capabilities.

D : Develop a line chart with CTR on the x-axis and date on the y-axis, and add a filter for campaign in the dashboard to isolate specific campaigns.

Answer: B

Question 3

You have a model registered in the Model Registry in Vertex AI and need to evaluate its performance over multiple versions for A/B testing in production. Which approach will best allow you to manage different versions effectively while ensuring they can be accessed and deployed as needed?

Options :

A : Use Vertex AI Experimentation instead of Model Registry for version management.

B : Override the model version in Model Registry each time you retrain to avoid clutter.

C : Assign a unique alias to each model version in Model Registry for easier version tracking and deployment.

D : Manually track all versions in a local database and deploy them to a staging environment one by one.

Answer: C

Question 4

Your retail organization stores sensitive application usage data in Cloud Storage. You need to encrypt the data without the operational overhead of managing encryption keys. What should you do?

Options :

A : Use Google-managed encryption keys (GMEK).

B : Use customer-managed encryption keys (CMEK).

C : Use customer-supplied encryption keys (CSEK).

D : Use customer-supplied encryption keys (CSEK) for the sensitive data and customer-managed encryption keys (CMEK) for the less sensitive data.

Answer: A

Question 5

Your team is loading data into BigQuery from various sources, including CSV files, JSON logs, and a relational database. During ingestion, you notice mismatched schemas, inconsistent null handling, and duplicate records. What should you do to identify and resolve these issues before loading the data?

Options :

A : Use Cloud Composer to write custom Python scripts for data validation.

B : Rely on source system data validation rules to ensure consistency.

C : Use Dataprep to visually inspect, clean, and standardize the data.

D : Load the data into BigQuery directly and identify issues using error messages.

Answer: C

Confidently Practice Online with Free Associate-Data-Practitioner Exam Cram

Buying Options: