Google Professional Data Engineer Question 219
Select 3Google Cloud PlatformYou have been tasked with conducting data discovery for a new analytics project in Google Cloud. The goal is to identify the structure, quality, and relevance of the data stored in BigQuery. Which of the following actions should you take to ensure a thorough data discovery process?
- A
Use the BigQuery Data Profiler to analyze the data's distribution, null values, and anomalies.
- B
Write custom SQL queries to manually inspect sample data and validate schema consistency.
- C
Export the data from BigQuery to a local file system for offline analysis using third-party tools.
- D
Leverage Data Catalog to explore metadata, lineage, and assigned data classifications.
- E
Run a BigQuery ML model to predict missing values and fill data gaps.
Show answer and explanation
Correct answers: A, B, D
Explanation
Data discovery involves analyzing the structure, quality, and metadata of datasets to assess their suitability for the project. Tools like BigQuery Data Profiler and Data Catalog provide powerful capabilities for on-platform data exploration, while custom SQL queries allow for targeted inspection. Exporting data off-platform and predictive modeling are not appropriate for this phase of the data lifecycle.
- A. Correct.
This is correct. The BigQuery Data Profiler is specifically designed to help analyze data quality, including null value counts, data distributions, and potential anomalies, which are key for data discovery.
- B. Correct.
This is correct. Writing custom SQL queries is an effective way to explore specific aspects of the data, such as checking for schema consistency, identifying outliers, or sampling the data for quality checks.
- C. Incorrect.
This is incorrect. Exporting data for offline analysis is inefficient and increases the risk of data security issues. Google Cloud provides native tools like BigQuery Data Profiler and Data Catalog for on-platform discovery.
- D. Correct.
This is correct. Data Catalog allows you to explore metadata, lineage, and classifications, which are critical for understanding the context, sensitivity, and structure of the data.
- E. Incorrect.
This is incorrect. Running a BigQuery ML model to predict missing values is not a standard data discovery method. Filling data gaps is a data preparation task, not part of the initial discovery process.