Best Practices
Organize Credentials and Metadata
Document API Keys
Keep a secure record of your API keys, endpoints, and query IDs. This documentation can include versioning details, usage notes, and any dependencies on other services.
Example: Store credentials and metadata in a version-controlled repository (with restricted access) or a secure credential manager.
Label Data Sources Clearly
In Tableau, label your data sources with descriptive names that map back to their corresponding queries in Scuba. Doing so makes it easier to track how each dataset is built and quickly diagnose discrepancies.
Design Queries with the Final Story in Mind
Determine Required Dimensions and Aggregations
Before creating or selecting a query ID, decide which dimensions and measures are essential to the narrative you want to tell. Incorporate only the fields you actually need to keep dashboards performant and uncluttered.
Set Start/End Times Appropriately
When defining time ranges, be aware of whether your queries use a fixed time window (e.g., “January 1 to January 31”) or a relative time window (e.g., “last 7 days”). Align this choice with how frequently your data updates and what your audience needs to see.
Maintain Granularity for Flexibility
Send Data at the Finest Needed Granularity
Ensure that timestamps are at the most granular level required for your analysis. If you need to roll up daily data into weeks or months, that’s easily done in Tableau. However, you cannot break down monthly timestamps into daily or hourly data if the underlying information is missing.
Tip: If there’s a chance you’ll analyze data at a daily resolution in the future, include daily timestamps now to avoid re-ingesting or re-processing data later.
Keep Data Flows Transparent
Map Queries to Dashboards
Create a simple reference that shows which query IDs fuel each dashboard or report. This practice helps teams quickly trace data back to its Scuba origins and evaluate if the query accurately serves the intended business need.
Review and Update Regularly
Business requirements evolve, and so do your queries. Periodically review your Time Series Connector setup in both Scuba and Tableau. Remove or refactor outdated queries and sources to maintain clarity and performance.
Test for Efficiency and Accuracy
Validate Performance at Scale
Run tests on large datasets to confirm that the connector handles your volume and complexity. High-level aggregations might not reveal certain bottlenecks, so stress-testing at full scale is recommended.
Cross-Check Data Consistency
Compare Tableau visualizations with Scuba dashboards to ensure the data matches, especially with relative time windows or complex filters. This step helps pinpoint misconfigurations or discrepancies in query definitions.
Further Resources
Scuba Documentation
Visit docs.scuba.io for additional details on query construction, time-handling intricacies, and advanced configuration options.