title: EconKit 计量经济学分析工具 emoji: 📊 colorFrom: blue colorTo: indigo sdk: streamlit sdk_version: "1.32.0" app_file: app.py pinned: true license: MIT tags:
- econometrics
- panel-data
- did
- causal-inference
- streamlit
- python
All-in-one Econometrics Toolkit | 一站式计量经济学实证分析工具
Designed for Chinese economics & management students (undergraduate thesis / master's / PhD) | No coding required | One-click academic charts and PDF reports
专为中国经管类学生设计（本科毕设 / 硕博论文）| 无需写代码 | 一键生成学术级图表与 PDF 报告
🚀 Live Demo — ModelScope (China) · 🚀 Live Demo — HuggingFace (Global) · ☕ Support
Keywords: DID Difference-in-Differences | PSM Propensity Score Matching | RDD Regression Discontinuity | IV Instrumental Variables | Panel Fixed Effects | Dynamic Panel GMM | Mediation | Moderation
| Module | Methods |
|---|---|
| 🔵 Descriptive & Diagnostics | Descriptive stats, correlation matrix, normality test, VIF, heteroskedasticity, autocorrelation |
| 🟡 Baseline Regression | OLS, individual/time/two-way FE, RE, Hausman test, panel unit root |
| 🔴 Causal Inference | DID (parallel trend + 1000x placebo), staggered DID, PSM, RDD, IV/2SLS, dynamic panel GMM |
| 🟢 Robustness | Winsorize, Bootstrap, placebo test, sample exclusion |
| 🟣 Heterogeneity & Mechanism | Subgroup regression, quantile regression, mediation (Bootstrap), moderation |
🤖 Smart Recommendation: Describe your research background, get a matched analysis path automatically.
- 🤖 ModelScope (China): https://modelscope.cn/studios/JackyCufe/EconKit/summary
- 🤗 HuggingFace (Global): https://huggingface.co/spaces/JackyCufe/econkit
git clone https://github.com/JackyCufe/econkit.git
cd econkit
pip install -r requirements.txt
streamlit run app.pyOpen http://localhost:8501 in your browser.
Supports Excel / CSV, panel data in long format:
| firm_id | year | treat | post | did | tfp | size | lev |
|---|---|---|---|---|---|---|---|
| 1 | 2010 | 0 | 0 | 0 | 2.15 | 10.23 | 0.32 |
| 2 | 2015 | 1 | 1 | 1 | 2.58 | 11.30 | 0.43 |
Unsure about format? Built-in sample data lets you explore the full workflow first.
EconKit uses academically validated Python libraries:
| Library | Purpose |
|---|---|
statsmodels |
OLS, panel regression, time series |
linearmodels |
FE, RE, IV/2SLS |
econml |
DID, causal inference |
scikit-learn |
PSM propensity score estimation |
rdrobust |
RDD |
Replication verified: Classic DID results reproduced with publicly available panel data, coefficient error within 0.01. For critical conclusions, consult a professional econometrician.
If EconKit helped you, consider buying me a coffee ☕
A ⭐ Star also helps more students discover this tool!
- v1.0.0 (2026-03-14) — MVP release, full econometrics method suite