- Home
- Skills
- Data & Databases
- Data Warehouse Architect — Build a Reliable Analytics
Works with the AI Agents you already use
Data Warehouse Architect — Build a Reliable Analytics
Design governed analytics foundations with explicit grain, history, recoverable loads, reconciliation, security and workload planning.
$9.99
Data Warehouse Architect — Build a Reliable Analytics
Example session with this skill installed
Illustrative scenario: Propose an analytics foundation for an online store using supplied daily order and customer exports. Finance needs daily net revenue by order date and region. Refund data and currency definitions are not supplied. No platform, volume, concurrency or freshness agreement exists. Provide model boundaries and verification gates without inventing schemas or selecting a vendor.
- Read your context and instructions
- Compiled the data warehouse architect
Warehouse architecture proposal — illustrative, platform-neutral
Decision scope: daily net revenue by order date and region. Required unknowns: authoritative order/customer keys, refund handling, currency and tax semantics, region history, source delivery, volume, query concurrency, retention and recovery objectives. Do not calculate net revenue before definitions are approved.
Proposed layers: immutable/replayable source landing where policy permits; staging preserving source fields and load metadata; integrated authoritative identities and approved history; order-level analytical model at a confirmed grain; reporting mart/semantic metric for net revenue. These are design proposals, not discovered source schemas.
Grain gate: confirm whether each source row is an order, line item or event. Model facts and joins accordingly to prevent double-counting. Define refund grain and attribution to order date separately.
History gate: decide whether region reporting uses purchase-time region or current customer region. Preserve change history if the chosen business question requires it; do not silently rewrite historical truth.
Loading: deterministic keys/version rules, idempotent incremental processing, quarantine of invalid/conflicting records and a documented replay/backfill path. Platform-specific syntax deferred.
Quality/reconciliation: confirm source totals, accepted/rejected records, refunds and currency control totals; test join fanout and metric definitions against authoritative finance evidence.
Operations: assign proposed data/metric owners; define access and masking, freshness observability, workload isolation, recovery testing and cost monitoring after real requirements are known.
Execution notes: no warehouse created, data loaded, benchmark run or savings measured. Reference design is not proof of production performance.
Connects securely to your tools. The creator never sees your data.
About this skill
Data Warehouse Architect designs and reviews analytical foundations spanning ingestion, raw/staging, integration, core models, marts and semantic serving. It starts with business decisions, source authority, grain, history, freshness and workload evidence while staying vendor-neutral unless a platform is specified.
The WAREHOUSE-96 workflow considers dimensional or other appropriate models, incremental loading, schema evolution, reconciliation, lineage, workload isolation, privacy, recovery and cost. It treats architectural patterns as choices and prefers reversible decisions under uncertainty. A completed load does not prove correct analytical data.
Use supplied schemas, sanitized samples, query plans and requirements. Without platform access it produces defensible designs and validation steps rather than claiming implementation. Exact capabilities and performance depend on real platforms and workloads; benchmark and recovery validation remain necessary.
Architecture depends on platform versions, workloads, distributions, team capability, obligations and commercial terms. Reference designs do not prove performance. Benchmarks and recovery need representative validation; no schemas, capabilities, savings or certifications are invented.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before any payment.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
Fresh listing
Recently published to Agensi
Money-back guarantee
Doesn't do what its page promises? You get your money back.
Trust & safety
Security scan
Security scan passed 1 month ago
Permissions
File Scopes
Use authorized metadata, requirements and masked samples. Optional read/write files support design and reports. Minimize sensitive rows and never expose secrets or endpoints. Database queries, cloud configuration and migration/deployment require explicit authorization and suitable tools.