Antigravity Skills - Build and Discover AI Agent SkillsAntigravity Skills

Data Analyst Skill · Data Ai

Data Analyst: From Raw Data to Actionable Insights

Transform data into actionable insights with rigorous statistical methodology and compelling narratives.

From trend detection to executive presentations.

  • Statistics
  • Machine Learning
  • Data Visualization
  • Business Intelligence
  • A/B Testing

About This Skill

Get instant access to a comprehensive toolkit for data analysis—from choosing the right statistical test to presenting findings that drive executive decisions.

Start Analyzing in Seconds

  1. 1Describe your data analysis challenge
  2. 2Get recommended statistical or ML methods
  3. 3Design metrics and thresholds with guidance
  4. 4Structure findings as compelling narratives

Core Capabilities

Statistical Method Selection

Choose the right test for your data: t-tests, ANOVA, regression, clustering, and more with decision trees and best practices.

A/B Testing & Experiments

Design statistically sound experiments, calculate sample sizes, interpret results with effect sizes and confidence intervals.

Trend & Anomaly Detection

Detect growth patterns, change points, and outliers using time series analysis, stationarity tests, and threshold design.

Attribution Modeling

Understand what drives conversions with first-touch, last-touch, linear, position-based, and data-driven attribution models.

User Segmentation

Group users effectively using K-means, hierarchical clustering, and classification with proper validation techniques.

Data Storytelling

Transform analysis into compelling narratives with problem-solution frameworks, visualization techniques, and executive templates.

SKILL.md

---
name: data-analyst
description: Use when analyzing data for business decisions, designing statistical or ML algorithms, creating data visualizations, or presenting insights to stakeholders. For trend detection, attribution modeling, A/B testing, user segmentation, and executive reporting.
---

# Data Analyst

Transform data into actionable insights and compelling narratives that drive business decisions.

## Overview

This skill combines rigorous statistical methodology with effective data storytelling. It helps you:
1. **Analyze**: Choose and apply the right statistical/ML methods
2. **Design**: Create thresholds, metrics, and decision logic
3. **Communicate**: Present insights that inspire action

## When to Use This Skill

- **Algorithm Design**: Designing data analysis algorithms or judgment logic
- **Method Selection**: Choosing between statistical tests, regression, or ML methods
- **Threshold Design**: Setting decision boundaries or classification criteria
- **Trend Analysis**: Detecting changes, growth patterns, or anomalies
- **Attribution**: Understanding what drives conversion or outcomes
- **Executive Reporting**: Presenting analytics to stakeholders
- **A/B Testing**: Designing and interpreting experiments

## Workflow

```
Problem → Classify → Select Method → Design Metrics → Set Thresholds → Tell Story
```

### Core Methodology

**Trend Detection**:
- Growth ratio = current_mean / baseline_mean
- CV (Coefficient of Variation) for stability
- Change point detection for inflection

**Statistical Comparison**:
| Scenario | Normal Data | Non-Normal |
|----------|-------------|------------|
| 2 groups | t-test | Mann-Whitney |
| 3+ groups | ANOVA | Kruskal-Wallis |

**Attribution Models**:
- First-touch (awareness)
- Last-touch (conversion)
- Linear (balanced)
- Position-based (40/20/40)
- Data-driven (ML)

### Reference Files

| File | Content |
|------|---------|
| `method_selection.md` | Decision tree for all analysis types |
| `time_series.md` | St

Frequently Asked Questions

FAQ

What's the difference between correlation and causation?
Correlation shows variables move together; causation means one actually causes changes in another. Use regression with controls, A/B tests, or causal inference methods to establish causation.
How do I choose between ML and statistical methods?
Use statistics when testing hypotheses, understanding 'why', or with small samples. Use ML when discovering patterns, predicting 'what', with large datasets, or when accuracy matters more than interpretability.
What should I report beyond p-values?
Always report effect sizes (Cohen's d, odds ratios), confidence intervals, and sample sizes. Statistical significance doesn't mean practical significance—a tiny effect can be 'significant' with enough data.

Discussion

Discussion

0 comments
U

On This Page

Trigger Phrases

Use these phrases to activate this skill in your AI coding assistant:

Analyze this dataWhat statistical test should I useDetect if trendingCompare these groupsAttribution modelSegment usersPresent to executivesDesign thresholds