What's changed: In-scope coverage: ops platform (ECS/EKS/CloudFormation/CDK/CLI/Code trio/SSM/Managed Grafana/Well-Architected/Cost Explorer/Budgets) + data-prep/ML (Glue DataBrew/SageMaker AI/Bedrock/Kendra)
3.2Analysis and Visualization
Understand turning data into value: SQL analytics with Athena/Redshift, BI dashboards with Amazon QuickSight, and log/search analytics with OpenSearch.
Stored data becomes valuable only when analyzed and visualized. Query with Athena/Redshift, visualize with QuickSight, and do log/search analytics with OpenSearch.
3.2.1Analyze, visualize, and search
- Athena/Redshift: analyze data with SQL (serverless S3 query / large DWH).
- Amazon QuickSight: visualize data in BI dashboards for business users (fast via SPICE).
- OpenSearch Service: strong for log analytics and full-text search, with dashboards (Kibana).
Common on DEA: business BI dashboards = QuickSight, log analytics/full-text search = OpenSearch, ad-hoc SQL on S3 = Athena. Don’t swap the roles of visualization and analysis.
Separate "analysis (query)," "visualization (dashboards)," and "search/log analytics." Query with Athena (serverless SQL on S3, billed per scan) or Redshift (large DWH). Visualize with Amazon QuickSight, sharing BI dashboards with business users; loading into SPICE (in-memory) speeds responses and offloads the source (with row/column-level security, natural-language Q, and embedding). For search and log analytics/observability, Amazon OpenSearch Service excels, visualized in OpenSearch Dashboards (Kibana). Chain the layers—"QuickSight uses Athena/Redshift/RDS as data sources," "ship logs into OpenSearch to search." The axes: "standardized business BI = QuickSight," "ad-hoc exploratory SQL = Athena," "full-text/log observability = OpenSearch." A common trap is swapping the visualization tool (QuickSight) with the analysis engine (Athena/Redshift).
| Goal | Use |
|---|---|
| Business BI dashboards | QuickSight (SPICE) |
| Ad-hoc SQL on S3 | Athena |
| Steady large-scale analytics | Redshift |
| Full-text/log analytics | OpenSearch Service |
Scenario: build a revenue dashboard for execs and a log-investigation platform for SREs. Use Redshift (or S3 + Athena) as a QuickSight data source, accelerate with SPICE, and segment by department with row-level security. Aggregate app/infra logs in OpenSearch and provide full-text search and observability dashboards via Kibana.
Q. Business BI dashboards? QuickSight. Q. Log analytics/full-text search? OpenSearch. Q. Ad-hoc SQL on S3? Athena. Q. Faster dashboards? Load into SPICE. Q. Segment by department? QuickSight row/column-level security.
Watch the mix-ups: (1) QuickSight visualizes; Athena/Redshift are the analysis engines—don’t swap (QuickSight alone isn’t a query engine). (2) OpenSearch is for full-text/logs, not the lead for large aggregation analytics (that’s Redshift). (3) SPICE has capacity limits—use direct query for very large data.
Loading data into QuickSight’s SPICE (in-memory engine) speeds dashboards and reduces load on the source.
3.2.2In-scope services for data prep, ML, and search
Before and after analysis, data-prep and machine-learning services are also tested. AWS Glue DataBrew cleans and normalizes data (250+ transforms) via a visual, no-code interface, speeding data preparation. To run ML on prepared features, Amazon SageMaker AI is a platform integrating data prep, training, deployment, and inference, bridging data engineering and ML via feature store and pipelines. To embed generative AI in a pipeline, Amazon Bedrock provides multi-provider foundation models via one serverless API, invocable from Lambda/Step Functions to summarize or classify extracted data. For natural-language search (question answering) over internal documents, use ML-powered Amazon Kendra, keeping its role distinct from SQL analytics on structured data (Athena) and full-text/log search (OpenSearch).
3.2.3Section summary
- Athena/Redshift (analyze) / QuickSight (visualize) / OpenSearch (search·logs)
- BI dashboards = QuickSight; full-text/logs = OpenSearch
- Data prep = Glue DataBrew; ML = SageMaker AI; gen-AI = Bedrock; enterprise search = Kendra
Sign in to track progress — Log in.
Quick check
(just a quick review)Q1. You want interactive BI dashboards to visualize data for business users. Which fits best?
Q2. Which service best fits full-text log search and observability analytics/dashboards?
Q3. You want serverless ad-hoc SQL queries on S3 data. Which fits best?