Skip to content

brycewang-stanford/Auto-Empirical-Research-Skills / skills/14-luischanci-claude-code-research-starter/dot-claude/skills/data-analysis

data-analysis

Verified

End-to-end R data analysis workflow from exploration through regression to publication-ready tables and figures.

FileKindSize
SKILL.mdskill md2.4 KB
---
name: data-analysis
description: End-to-end R data analysis workflow from exploration through regression to publication-ready tables and figures.
disable-model-invocation: true
argument-hint: "[dataset path or description of analysis goal]"
allowed-tools: ["Read", "Grep", "Glob", "Write", "Edit", "Bash", "Task"]
---

## Data Analysis

Run an end-to-end data analysis in R through 5 phases.

### Phase 1: Setup and Data Loading
- Read R code conventions from `.claude/rules/r-code-conventions.md`
- Create script with proper header (title, author, purpose, inputs, outputs)
- Load packages with `library()` (never `require()`)
- Set `set.seed()` if stochastic elements are involved
- Load data using relative paths

### Phase 2: Exploratory Data Analysis
- Summary statistics (mean, sd, min, max, N)
- Missingness patterns
- Distributions of key variables
- Bivariate relationships
- Time patterns (if panel data)
- Group comparisons
- Save diagnostics to `output/diagnostics/`

### Phase 3: Main Analysis
- Use `fixest` for panel data with fixed effects (see `/econometrics-r` for detailed patterns)
- Use `lm`/`glm` for cross-section
- Cluster standard errors appropriately
- Run multiple specifications (simple to complex)
- Report both statistical and economic significance
- For causal inference (IV, DiD, RDD), see `/causal-inference-r` for diagnostics and modern estimators
- For spatial econometrics, see `/econometrics-r` for `spdep`/`spatialreg` patterns

### Phase 4: Publication-Ready Output
- **Tables:** `modelsummary` (preferred) or `stargazer`, export as `.tex` and `.html`
- **Figures:** `ggplot2` with project theme, explicit dimensions, white background, 300 DPI
- Export figures as `.pdf` and `.png`

### Phase 5: Save and Review
- `saveRDS()` for all computed objects
- Run r-reviewer agent on generated script
- Address Critical/High issues

### Script Structure Template
```r
# ==============================================================================
# Title: [Analysis Name]
# Author: [Name]
# Date: [YYYY-MM-DD]
# Purpose: [What this script does]
# Inputs: [Data files]
# Outputs: [Tables, figures, RDS files]
# ==============================================================================

# 0. Setup ----
library(tidyverse)
library(fixest)
library(modelsummary)

# 1. Data Loading ----

# 2. Exploratory Analysis ----

# 3. Main Analysis ----

# 4. Tables and Figures ----

# 5. Export ----
```

All scripts saved to `scripts/R/`, all outputs to `output/`.
planning-and-task-breakdownaddyosmani/agent-skillsBreaks work into ordered tasks. Use when you have a spec or clear requirements and need to break work into implementable tasks. Use when a task feels too large to start, when you need to estimate scope, or when parallel work is possible.Stars: 119kUpdated Claude CodeCodexVerifiedInstall
screenpipethedotmack/claude-memSearch screen and audio history recorded by Screenpipe and optionally save selected context into claude-mem. Use when the user asks to recall desktop activity, a meeting, or something seen in another app through Screenpipe.Stars: 118kUpdated Claude CodeCodexVerifiedInstall
contact-widgetnexu-io/open-designSelf-contained floating chat widget with welcome screen, social links, meeting button, and message input. Single HTML file, zero dependencies.Stars: 108kUpdated Claude CodeCodexVerifiedInstall
meeting-notesnexu-io/open-designMeeting notes page — title bar with attendees, agenda checklist, decisions block, action items table with owners + dates, and a "next meeting" footer. Use when the brief mentions "meeting notes", "minutes", "1:1 notes", "all-hands recap", or "会议纪要".Stars: 108kUpdated Claude CodeCodexVerifiedInstall