Data Analyst

What it does

This skill enables an AI agent to function as a senior data analyst, performing tasks such as writing production SQL queries, designing visualizations, running statistical tests (t-tests, chi-square), and translating findings into actionable business recommendations. It focuses on framing business questions as testable hypotheses, exploring datasets for anomalies, and delivering insights in a structured "What / So What / Now What" format. The agent will use CTEs to improve SQL readability and follows best practices like filtering early and aggregating late.

When to use it

• Analyzing campaign performance to determine if specific campaigns impacted key metrics (e.g., 7-day retention).
• Understanding customer behavior through cohort analysis or funnel analysis.
• Comparing groups of data using statistical hypothesis testing.
• Identifying relationships between variables using regression analysis.
• Creating visualizations to communicate data insights effectively.

Key capabilities

• SQL Query Generation: Writes production-ready SQL queries, including the use of Common Table Expressions (CTEs).
• Data Exploration & Profiling: Calculates descriptive statistics and checks for nulls, duplicates, and outliers.
• Statistical Analysis: Performs cohort analysis, funnel analysis, t-tests, chi-square tests, and regression.
• Visualization Design: Selects appropriate chart types (line, bar, donut, histogram, scatter) and adheres to design rules (e.g., Y-axis at zero for bars).
• Insight Delivery: Structures findings using the "What / So What / Now What" framework.

Example prompts

• “Did campaign X increase 7-day retention by >= 5%? Please provide a visualization.”
• “Perform a cohort analysis to understand customer retention trends over time.”
• “Analyze order data and identify any correlations between order date and amount."

Tips & gotchas

• The agent requires access to relevant datasets and the ability to execute SQL queries.
• Be specific with your business questions to ensure the agent can formulate a testable hypothesis.
• Familiarity with basic statistical concepts will help you interpret the agent’s findings.