python-econ-computing
Use when writing Python code for DSGE models, HANK models, numerical economic computation, causal inference, or quantitative economic data analysis
Profile →npx skills add brycewang-stanford/Auto-Empirical-Research-Skills --skill 20-wenddymacro-python-econ-skill --agent claude-code
Same command for any agent — swap --agent for codex, cursor, copilot.
Weekly change comes from our own snapshots, not the repository page — it measures attention, not adoption.
What it does
Guides the agent to write Python code for macroeconomic models, causal inference, and data analysis, emphasizing vectorization first, using Numba to accelerate loops, and aligning code structure with economic theory.
How it works
Outlines recommended libraries for numerical core (numpy, scipy), loop acceleration (numba with @njit, or parallel=True), and economics toolkits (quantecon, HARK, sequence_jacobian). Specifies usage for linear DSGE solution (Blanchard-Kahn), perturbation methods, and steady-state solving. Provides examples for HANK with Sequence-Space Jacobian, and a NumPy/Numba VFI skeleton. Recommends pyfixest for fixed-effects linear models and diff-diff for DiD/DD/DDD workflows, including how to run estimators, specify data, controls, FE, clustering, and inference.
Describes DID/DDD workflows stepwise, including data prep, parallel trends checks, multiple regression specifications (M1–M6 with unit and time FE, covariates, and various fixed effects), event studies, and robustness checks (placebo tests, Goodman-Bacon, subsamples). Includes EventStudy usage, interaction-based heterogeneity, and IV/2SLS/GMM/LIML machinery with linearmodels, plus diagnostics like first-stage F-statistics and Wooldridge tests.
When to use it
Use when writing Python for DSGE/HANK modeling, numerical economic computation, causal inference analyses, or data analysis tasks as described in the overview.
What it can touch
Libraries and tools mentioned include: numpy, scipy, numba, quantecon, sequence_jacobian, HARK, pyfixest, diff-diff, pandas, matplotlib, seaborn, pysynth, synth_control, sdid, causalml, pymatch, econml, dowhy, rdrobust, rddensity, rdlocrand, linearmodels, and various did/did workflows. Specific commands shown include: import numpy as np, from scipy.linalg import ordqz, @njit, @sj.simple, model.solve_steady_state, pf.feols, diff_diff.DiD, EventStudy, CS, SA, BJS, DDD, pf.iplot, res.plot, and IV2SLS/IVGMM/IVLIML usage."
Caveats
The skill references multiple external libraries and assumes availability via pip install commands, but does not state licensing beyond NOASSERTION. It provides example code and expects correct context-specific calibration and data structures; no guarantees of results or stability across environments are given.
# Python Economic Numerical Computing - Author：Wenli Xu - Email： wlxu@cityu.edu.mo - 2026-03-11 --- ## Overview Best practices for macroeconomic modeling (DSGE/HANK), causal inference, and data analysis in Python. Core principle: **vectorize first, accelerate loops with Numba, keep code structure aligned with economic theory**. --- ## Library Quick Reference | Use Case | Preferred Libraries | |---
What does the python-econ-computing skill do?
Use when writing Python code for DSGE models, HANK models, numerical economic computation, causal inference, or quantitative economic data analysis
How do I install it?
Run `npx skills add brycewang-stanford/Auto-Empirical-Research-Skills --skill 20-wenddymacro-python-econ-skill --agent claude-code` — it drops the skill into your project so the agent can pick it up. Swap the --agent value for codex, cursor or copilot if you use one of those.
Where does this skill come from?
From brycewang-stanford/Auto-Empirical-Research-Skills, a repository with 3,244 stars. We read it straight from the repository tree rather than a submitted listing, so what you see here is what is actually published.
Is a popular skill a good skill?
Not necessarily. Stars measure attention, not adoption — a repository can trend for a week and be abandoned. That is why we show the weekly change from our own snapshots next to the total, instead of a single flattering number.