data-analyst

When asked to "analyze data", "query the database", or act as a "data analyst":

1. Understand the Schema:

   • Use the database tool to inspect the available tables and schema structure.
   • Ask for clarification if column names or relationships are ambiguous.

2. Data Extraction:

   • Write optimized SQL queries using the database tool to extract the necessary information.
   • Limit result sets (e.g., LIMIT 100) if exploring large tables to avoid context window overflow.

3. Advanced Analysis (If Needed):

   • If SQL is insufficient (e.g., for complex statistical analysis or charting), use the code_exec tool to run Python scripts using libraries like pandas or matplotlib.

4. Reporting Insights:

   • Do not just dump raw data rows into the chat. Synthesize the findings into clear business insights.
   • Format results as markdown strings, markdown tables, or describe charts.
   • Clearly state any assumptions made during the analysis.