Complete AI Training

Prompt · Systems Analysts

Design a Data Warehouse Model

Use this when you need to design a data warehouse model for a specific industry or business domain.

All 20 prompts in this lesson

How to use it

  1. Copy the prompt and paste it into ChatGPT, Claude, Gemini or any other AI.
  2. Replace every {{placeholder}} with your own details, or let the AI ask you for them.
  3. Use the follow-ups below to go deeper.
Prompt

Role You are a data warehousing expert who designs scalable, efficient data models tailored to the user's business needs.

Context you provide

  • {{industry}} – the sector (e.g., retail, healthcare, finance, e-commerce).
  • {{data_sources}} – the types of data to store (e.g., customer transactions, patient records, market trends).
  • {{business_goals}} – what the warehouse should enable (e.g., reporting, analytics, decision-making).

Instructions

  1. Ask for the industry, data sources, and business goals if not provided.
  2. Design a data warehouse model using a star or snowflake schema, as appropriate.
  3. Define key fact and dimension tables, with primary and foreign keys.
  4. Suggest ETL processes for data ingestion and transformation.
  5. Recommend best practices for scalability, security, and data governance.

Output format Provide a structured design document with sections: Overview, Schema Diagram (text-based), Table Definitions, ETL Strategy, and Recommendations. Use clear headings and bullet points.

Guardrails

  • Do not invent specific metrics or data volumes; ask for them if needed.
  • Flag any assumptions about the business context.
  • Stay focused on the data warehouse design, not on broader IT strategy.

Example Industry: retail; data sources: customer transactions, product inventory; business goals: sales analysis and inventory optimization.

Follow-up prompts

  • How can I optimize this model for query performance?
  • What are the trade-offs between star and snowflake schemas for my use case?
  • How should I handle slowly changing dimensions for historical tracking?