Data Operations and Support
Query, Analysis, and Visualization
CorePrepare, verify, query, aggregate, and visualize governed data using the engine and capacity model appropriate to its location and workload.
Aligned to AWS Certified Data Engineer - Associate (DEA-C01) Version 1.1, verified August 25, 2026.
Why this matters
Technically useful analysis depends on valid inputs, correct grain, and a suitable execution model. A chart cannot prove quality, and a shared SQL language does not make all engines operationally equivalent.
Must Know
- Use current Amazon Quick surfaces to publish governed datasets and visualizations with refresh and access configured; visualization remains separate from data validation.
- Profile and clean nulls, duplicates, invalid categories, and type errors through explicit rules while retaining rejected-record evidence.
- Use Athena SQL for cataloged S3 data and serverless queries; use Redshift SQL and views for warehouse-local joins and repeated analytical workloads.
- Use Athena notebooks with Apache Spark for interactive distributed exploration when SQL alone is insufficient.
- Compare provisioned and serverless models on demand shape, startup behavior, tuning control, utilization, and operating effort—not list price alone.
- Define the output grain before choosing grouping, pivoting, or a window such as rolling average.
- Deduplicate and validate data before aggregation so technical metrics are not biased by invalid records.
Compare and Distinguish
- Amazon Quick visualization versus data-quality rules: presentation differs from validation.
- Athena SQL versus Redshift SQL: cataloged S3 serverless access differs from warehouse-local optimized processing.
- Athena SQL versus Athena Spark notebook: declarative query differs from interactive distributed exploration.
- Provisioned versus serverless: explicit capacity and tuning differ from demand-driven capacity.
Scenario examples
- An engineer queries partitioned S3 logs with Athena and uses Redshift for repeated warehouse joins.
- An Athena Spark notebook explores cataloged data interactively while a production view remains SQL-managed.
- Duplicate events are removed before a per-tenant rolling average is calculated.
Exam traps
- Assuming a visually plausible dashboard proves completeness and uniqueness.
- Choosing serverless for every workload without considering sustained utilization or tuning.
- Computing a rolling average without partition and order semantics.
- Using a notebook when a governed repeatable SQL view is the actual requirement.
Key takeaways
- Validate before visualizing or aggregating.
- Choose the engine where the governed data and workload fit.
- Separate capacity model from query language.
How it works
- The engineer validates and prepares the source, executes the query in the engine suited to its location, and publishes only governed results.
- Windowed calculations define partition, ordering, frame, and output grain before visualization or reuse.
When to use it
- Use Athena SQL for cataloged S3 data, Redshift SQL for repeated warehouse processing, and Athena Spark notebooks for interactive distributed exploration.
- Use Amazon Quick to publish refreshable governed views after the source and query outputs have been validated.
Security and governance implications
- Authorize the source, query-result location, notebook, dataset, and visualization audience as separate resources.
- Prevent technical exploration from exposing sensitive data or turning an ungoverned notebook result into a production contract.
Common failure modes and diagnosis
- For an incorrect aggregate, inspect duplicates, nulls, filters, grouping grain, window partition, order, and frame.
- For slow or stale results, compare data location, partitions, warehouse workload, refresh status, and the chosen capacity model.
More detail
- 3.2.1: Visualize data by using AWS services and tools such as AWS Glue DataBrew and Amazon Quick.
- 3.2.2: Verify and clean data by using appropriate preparation, query, notebook, or transformation tools.
- 3.2.3: Use SQL in Amazon Redshift and Athena to query data or to create views.
- 3.2.4: Use Athena notebooks that use Apache Spark to explore data.
- 3.2.5: Describe tradeoffs between provisioned services and serverless services.
- 3.2.6: Define data aggregation, rolling average, grouping, and pivoting.
Ready for the quiz?
- Is the governed data best queried in S3, processed in the warehouse, or explored interactively with Spark?
- What grain, partition, order, frame, and quality checks make the requested aggregation technically correct?
Related objectives
- D3.2 — Task 3.2: Analyze data by using AWS services
- 3.2.1 — Visualize data by using AWS services and tools (for example, DataBrew, Amazon QuickSight).
- 3.2.2 — Verify and clean data (for example, Lambda, Athena, QuickSight, Jupyter Notebooks, Amazon SageMaker Data Wrangler).
- 3.2.3 — Use SQL in Amazon Redshift and Athena to query data or to create views.
- 3.2.4 — Use Athena notebooks that use Apache Spark to explore data.
- 3.2.5 — Describe tradeoffs between provisioned services and serverless services.
- 3.2.6 — Define data aggregation, rolling average, grouping, and pivoting.