pyfixest-latex

PyFixest LaTeX Generator

7 functions for converting PyFixest models to publication-quality LaTeX tables and figures.

Tables (4): regression, dynamic/event study, robustness (grouped), summary statistics Figures (3): event study plot, treatment assignment heatmap, coefficient comparison

Setup

Option A: Run setup script

python <skill-path>/scripts/setup_module.py /path/to/project

Option B: Manual copy

1. Copy assets/pyfixest_latex/ to your project root
2. Install dependencies: pip install pyfixest pandas numpy scipy matplotlib
3. Create output dirs: output/tables/ and output/figures/

Configure paths (required before generating output)

from pyfixest_latex import set_output_path, set_figure_output_path
set_output_path("Results/Tables")           # or any absolute/relative path
set_figure_output_path("Results/Figures")   # or any absolute/relative path

Core Workflow

import pyfixest as pf
from pyfixest_latex import (
    set_output_path, set_figure_output_path,
    create_regression_table, create_event_study_plot,
    create_summary_statistics_table
)

# 1. Configure output paths
set_output_path("Results/Tables")
set_figure_output_path("Results/Figures")

# 2. Fit models
m1 = pf.feols("Y ~ treat | unit + year", data, vcov={"CRV1": "unit"})
m2 = pf.feols("Y ~ treat + X1 | unit + year", data, vcov={"CRV1": "unit"})

# 3. Generate table
create_regression_table(
    models=[m1, m2],
    model_names=["Basic", "Controls"],
    title="Treatment Effects",
    label="tab:main",
    variable_labels={"treat": "Treatment Effect", "X1": "Control Variable"},
    depvar_labels={"Y": "Outcome Variable"},
    felabels={"unit": "Unit FE", "year": "Year FE"}
)
# → Saves: Results/Tables/main_regression.tex

Function Quick Reference

Tables

Function	Purpose	Key Params
create_regression_table()	Multi-model comparison	models, model_names, title, label
create_dynamic_table()	Event study coefficients	models, time_var, treat_var, reference_period, treatment_period
create_robustness_table()	Grouped specifications	model_groups, group_names
create_summary_statistics_table()	Descriptive stats	data (DataFrame), variables, variable_labels

Figures

Function	Purpose	Key Params
create_event_study_plot()	Dynamic effects plot	model, style (errorbar/filled/step), confidence_level
create_treatment_assignment_plot()	Treatment timing heatmap	data, unit, time, treat
create_coefficient_comparison_plot()	Forest plot across specs	models, model_names, use_ci

Utilities

Function	Purpose
set_output_path(path)	Configure table output directory
set_figure_output_path(path)	Configure figure output directory
list_saved_tables()	List generated .tex files
list_saved_figures()	List generated .png/.pdf files

Full API: See references/tables-api.md and references/figures-api.md

Example Templates & Assets

Complete DiD Analysis Template (849 lines):

• assets/example_did_analysis_template.py - Production-ready DiD analysis with DidConfig class
• Quick start: Copy, customize 7 settings in DidConfig, run
• See references/did-template-quick-reference.md for 30-second setup
• See references/did-template-guide.md for comprehensive documentation

Modular Function Examples:

• assets/1---example_summary_statistics.py - Standalone summary stats demonstration
• assets/2---example_figure_usage.py - Standalone figure generation examples
• assets/3---example_enhanced_table_generator.py - Standalone table generation examples
• assets/4---run_all_examples.py - Orchestrator to run all modular examples

Which to use?

• DiD template: Copy for complete analysis workflow (data → tables → figures)
• Modular examples: Learn individual functions in isolation, mix-and-match

Reference Documentation

Use these guides to learn specific patterns and troubleshoot issues:

Starting Out?

• Read workflow-patterns.md for complete analysis structure, UTF-8 setup, and best practices

Preparing Data?

• Read data-requirements.md for DataFrame structure, panel data setup, and validation

Need Usage Examples?

• Read common-patterns.md for 10 complete patterns (DiD, event studies, robustness checks, multiple outcomes)

DiD Template Documentation?

• Read did-template-quick-reference.md for 30-second DiD setup
• Read did-template-guide.md for comprehensive DiD guide

Customizing Functions?

• Read tables-api.md for all table function parameters
• Read figures-api.md for plot customization (colors, styles, CI levels)

Troubleshooting?

• Read troubleshooting.md for common issues and solutions

Using Statsmodels Instead of PyFixest?

• Read stargazer-alternative.md

Critical Rules

1. Set output paths BEFORE calling any generation function — paths are resolved at call time
2. Use type="tex" in PyFixest — the module calls pf.etable(..., type="tex") internally
3. Pass treatment_period to create_dynamic_table() and create_event_study_plot() for correct relative-time labels ($t_{-2}$, $t_0$, $t_{+1}$)
4. Unit FE before Time FE — automatic reordering in all tables (recognizes 25+ FE naming patterns)
5. CI distribution — t-distribution for n<30, normal for n>=30 (proper small-sample inference)
6. Figure backend — uses matplotlib 'Agg' (non-interactive); no GUI required
7. Windows — UTF-8 encoding handled automatically in figure generator
8. Filenames — auto-generated from label parameter: {label}_regression.tex, {label}_dynamic.tex, etc.

LaTeX Integration

Required packages in preamble

\usepackage{booktabs}
\usepackage{threeparttable}
\usepackage{graphicx}
\usepackage{makecell}    % for multi-line cells with \makecell

Include generated outputs

% Tables
\input{Results/Tables/main_regression.tex}
\input{Results/Tables/event_study_dynamic.tex}

% Figures
\begin{figure}[ht!]
\centering
\includegraphics[width=0.9\textwidth]{Results/Figures/event_study.png}
\caption{Event Study: Treatment Effects Over Time}
\label{fig:event_study}
\end{figure}

Table output structure

All tables use this LaTeX structure:

\begin{table}[ht!]
\centering
\caption{\\Title}
\parbox{\linewidth}{\footnotesize Notes text here}
\label{tab:label}
\footnotesize
\begin{threeparttable}
\begin{tabular}{lccc}
...
\end{tabular}
\end{threeparttable}
\end{table}