Data Analysis Workflow
Run an end-to-end data analysis: load, explore, analyze, and produce publication-ready output.
Input: $ARGUMENTS — a dataset path, analysis goal, or language preference.
Language Selection
- If user says "stata" or "do-file" → write a .do file
- If user says "R" or "Rscript" → write an .R file
- If ambiguous → ask. Default to Stata for regressions, R for visualization.
Project Data
| Dataset | Path | Contents |
|---|---|---|
| Panel A | data_final/{{panel_A}}.dta | {{description}} |
| Panel B | data_final/{{panel_B}}.dta | {{description}} |
| Panel C | data_final/{{panel_C}}.dta | {{description}} |
| Results CSV | output/results/{{regression_results}}.csv | {{description}} |
| Elections | data_final/{{elections_panel}}.dta | {{description}} |
Workflow Phases
Phase 1: Setup
- Follow conventions in
.claude/rules/stata-r-conventions.md - Create script with proper header
- Load data and inspect
Phase 2: Exploratory Analysis
- Summary statistics, distributions, missingness
- Treatment/control comparisons
- Time trends
Phase 3: Main Analysis
- Stata:
aregorreghdfewith county+year FE,vce(cluster {{cluster_var}}) - R:
fixest::feols()withcluster = ~{{cluster_var}} - Start simple, progressively add controls
- Report standardized effects (β/ȳ and β/σ_w)
Phase 4: Publication-Ready Output
- Stata:
esttab→.textables - R:
modelsummary→.texand.htmltables - Figures:
ggplot2with explicit dimensions, export PDF+PNG
Phase 5: Verify
- Run the code and check for errors
- Cross-check against
{{regression_results}}.csvif applicable - Save all outputs to
output/
Important
- New analysis → new numbered do-file (09+) or script in
scripts/ - Never modify 01–08 do-files without explicit permission
- All paths relative to
$RB/ - Match existing specifications before extending
