/data-analyzer

name: data-analyzer
description: Guidance for analyzing structured data, generating statistics and producing data-driven insights. Use when the user asks to analyze data, compute statistics, find patterns, or generate analytical reports.
license: Apache-2.0
allowed-tools: data_extractor rule_matcher document_parser
metadata:
  version: 1.0.0
  category: data
  tags: [data, analysis, statistics, report, insights]

Data Analysis

You are assisting with data analysis tasks. Follow these guidelines.

Analysis Workflow

1. **Understand** the data: identify columns, types, ranges, and any quality issues. 2. **Clean** the data: handle missing values, outliers, and format inconsistencies. 3. **Analyze**: compute relevant statistics (counts, sums, averages, distributions). 4. **Compare**: when multiple datasets or time periods exist, provide comparative analysis. 5. **Summarize**: present findings clearly with key metrics highlighted.

Statistical Methods

• Use descriptive statistics (mean, median, mode, std dev) as a baseline.
• Identify trends and patterns — year-over-year, month-over-month, category breakdowns.
• Flag outliers and anomalies with context about their potential significance.
• For comparisons, compute both absolute and percentage differences.

Output Formats

• **Summary**: Concise paragraph with key findings and numbers.
• **Table**: Structured tabular format for detailed breakdowns.
• **Report**: Sectioned report with executive summary, methodology, findings, and recommendations.

Best Practices

• Always state the sample size and time range of the data being analyzed.
• Round numbers appropriately for readability (2 decimal places for percentages).
• When making comparisons, ensure the baseline and comparison period are clear.
• Distinguish between correlation and causation in findings.
• Provide actionable recommendations when the analysis supports them.