Data Analyst

---
name: Data Analyst
description: Analyze datasets to find patterns, outliers, distributions, and generate narrative insights
version: 1.0.0
author: chvor
type: workflow
category: data
icon: bar-chart
tags:
  - data-analysis
  - statistics
  - csv
  - json
  - insights
  - visualization
  - outliers
  - distributions
  - analytics
---
When the user shares data (CSV, JSON, tables) or asks for analysis:

## Process

1. **Understand the data**: What are the columns/fields? What does each row represent? What's the time range?
2. **Describe the basics**:
   - Row count, column count
   - Data types per column
   - Missing values (count and percentage)
   - Unique value counts for categorical columns
3. **Summarize distributions**:
   - Numerical: min, max, mean, median, std dev, quartiles
   - Categorical: top 5 values with counts
   - Temporal: date range, gaps, frequency
4. **Find patterns**:
   - Correlations between numerical columns
   - Trends over time
   - Grouping patterns (e.g., sales by region, errors by type)
5. **Spot anomalies**:
   - Outliers (values > 3 std devs or outside IQR)
   - Sudden changes in trends
   - Unexpected nulls or zeros
6. **Generate insights**: Write 3–5 narrative findings in plain English, not just numbers

## Output format

Start with a **1-sentence summary** of the dataset.

Then present findings as:
- **Finding**: [what you found]
- **Evidence**: [the numbers]
- **Implication**: [why it matters or what to investigate next]

## Visualization