5th-Stata-meeting-Slides.md · Evainquixotic/五分钟Markdown

---

marp: true

---

![bg right:57% brightness:. sepia:50%](https://fig-lianxh.oss-cn-shenzhen.aliyuncs.com/Stata_meeting-5th-graph-7c.png)

第五届 Stata 中国用户大会

#### 一个 Stata 用户的若干思考

###### 连玉君 (中山大学)
arlionn@163.com

&#x1F449; 幻灯片下载：<https://gitee.com/arlionn/md>

--- - --

### Stata 的前世今生

- [History of Stata](https://www.stata.com/support/faqs/resources/history-of-stata/)，`help version`
- Stata Press, 2015, [Thirty Years with Stata: A Retrospective](https://www.stata-press.com/books/tywsar-download.pdf)

--- - --

### 提纲

- 用户
- 老师
- Stata 公司

--- - --

## 1. 对用户的建议

- Stata 的角色
  - 基础
  - 资源
  - 习惯

--- - --

## Stata：16 &rarr; 17 想到的

- `help whatsnew16to17` &emsp;  [New in Stata17](https://www.stata.com/new-in-stata/)
- [Jupyter Notebook with Stata](https://www.stata.com/new-in-stata/jupyter-notebooks/)
- [pystata](https://www.stata.com/python/pystata/index.html)：[Example 1](https://www.stata.com/python/pystata/notebook/Example1.html)；[API]()
- Machine Learning - [H20.ai](https://docs.h2o.ai/) &rarr; [H2O integration](https://www.stata.com/new-in-stata/h2o-integration/)
- [Java integration](https://www.stata.com/new-in-stata/java-integration/) (Java: 电子交易系统；高频交易；安全)

> &emsp; &emsp;&emsp;&emsp; **Python** &emsp; **Java** 
> &emsp; 
> **Stata** &emsp; **Jupyter Notebook** &emsp; **R** 
> &emsp; 
> &emsp; **Matlab** &emsp; **C++** &emsp; **Fortran**

--- - --

### 学习/使用 Stata：我们的猎物

--- - --

### 学习 Stata

- 基本知识架构：OLS，GMM，MLE &rarr; 因果推断 &rarr; 机器学习？
- Conditional Expectation function (**CEF**)：
  - $E(y|\bold{x}) = f(\bold{x};\beta) = \alpha + x_{1}\beta_1 + x_{2}\beta_2 + \varepsilon$
  - Marginal Expectation Eﬀects (**MEE**)
  - **OLS** &rarr; Dummy &rarr; FE, RE, DID, RDD
  - **GMM** &rarr; IV, 2SLS, Dynamic Panel Data
- Conditional Probability Function (**CPF**)：
  - $P(y|\bold{x};\beta) = g(\bold{x};\beta)=\Phi(\bold{x}'\beta)$ or $g(\bold{x};\beta)=\frac{exp(\bold{x}'\beta)}{1+exp(\bold{x}'\beta)}$
  - Marginal Probability Eﬀects (**MPE**)
  - **MLE** &rarr; Probit/Logit, Count Data, Tobit, Heckman, Treament

--- - --

### 再看我们的猎物：CEF _v.s_ CPF

--- - --

### 经验总结

- 打基础：`Prabability`, `Matrix`, `Mata`, `Bayes`
  - **慢下来才能快起来**
  - 基本概念、核心原理 (不变) &rarr; 应万变 
    - 理解了 CPF &rarr; Probit, Tobit, Heckman, Count data
  - 煮肉：一次性煮熟 &emsp; Econometrica &emsp; 背后运球
- 细嚼慢咽 + 反刍
  - 一个主题 3-5 本书 &rarr; 整理笔记 
  - 读文献、干中学 ( What is Bootstrap ? )
  - 讲出来、写出来

--- - --

### 工具：协作+分享 | 资源

- [github](https://github.com/) &rarr; [my Github](https://github.com/arlionn?tab=repositories) &emsp; [github-Stata](https://github.com/search?l=Stata&o=desc&q=stata&s=stars&type=Repositories)
  - 团队协作：[worldbank/stata](https://github.com/worldbank/stata/tree/master/docs), [mostly-harmless](https://github.com/vikjam/mostly-harmless-replication), [stata-fundamentals](https://github.com/dlab-berkeley/stata-fundamentals) 
  - 代码分享：[reghdfe](https://github.com/sergiocorreia/reghdfe), [ivreghdfe](https://github.com/sergiocorreia/ivreghdfe), [moremata](https://github.com/benjann/moremata), &ensp;[gtools](https://github.com/mcaceresb/stata-gtools), [xtdcce2](https://github.com/JanDitzen/xtdcce2)
  - 教程和代码集合：[ML](https://github.com/trekhleb/homemade-machine-learning) (18.2k Star), [PythonDataScienceHandbook](https://github.com/jakevdp/PythonDataScienceHandbook) (30.5k Star)

- gitee ([码云](https://gitee.com/)) 
  - [五分钟 Markdown](https://gitee.com/arlionn/md) &emsp;  [synth](https://gitee.com/arlionn/collections/1257) &emsp; [论文复现-28篇](https://gitee.com/arlionn/collections/175903)
  - [lianxh](https://gitee.com/arlionn/lianxh), [songbl](https://gitee.com/songbolin/songbl)

--- - --

> Source: [ML](https://github.com/trekhleb/homemade-machine-learning) (18.2k Star)

&emsp; &emsp; &emsp; <img style="width: 850px" src="https://fig-lianxh.oss-cn-shenzhen.aliyuncs.com/20210819085530.png">

--- - --

### 工具：Github &rarr; Gitee &rarr; 星选集

- [我的](https://gitee.com/arlionn) 星选集：<https://gitee.com/arlionn/collections>

- [操作](https://gitee.com/help/articles/4261?utm_source==gitee-help-widget#article-header0)：**Fork** Github 仓库 &rarr; 进入 Gitee 
  &emsp; &emsp; &emsp; &rarr; 导入 Github 仓库 &rarr; **Star**, 拖入「星选集」

--- - --

--- - --

--- - --

### 笔记工具 - Markdown

- 五分钟 Markdown：<https://gitee.com/arlionn/md>, [新浪视频](https://weibo.com/tv/show/1034:4484204327796746?from=old_pc_videoshow)

![](https://fig-lianxh.oss-cn-shenzhen.aliyuncs.com/20210818124230.png)

--- - --

### 笔记工具 - Markdown

- Stata 17 也支持 Markdown 语法了
- Jupyter Notebook 主要用 Markdown 来码字
- [VScode](https://code.visualstudio.com/) + [Typora](https://www.typora.io/) + [web.marp.app](https://web.marp.app/)

--- - --

### 笔记工具 - LaTeX

- 剑桥大神 [Dexter Chua](https://dec41.user.srcf.net/) 的 [Cambridge Notes](https://dec41.user.srcf.net/notes/) 从大一到大四总计 4000 多页 
  - [github](https://github.com/dalcde/cam-notes) &rarr; [下载](https://github.com/dalcde/cam-notes/archive/refs/heads/master.zip)

![](https://fig-lianxh.oss-cn-shenzhen.aliyuncs.com/20210818122819.png)

--- - --

### 笔记工具 - LaTeX

> [TeX Live](http://tug.org/texlive/) + [TeX Studio](http://texstudio.sourceforge.net/) + [Mathpix](https://mathpix.com/ocr) &emsp; &emsp; [安装指南](https://zhuanlan.zhihu.com/p/41855480)

![](https://fig-lianxh.oss-cn-shenzhen.aliyuncs.com/Stata_meeting-5th-TeXstudio-01.png)

--- - --

>  **Help**：`texdoc pdfplots` (pp.68)   
> 4.3.8 Keys To Configure Plot Graphics  
> `\usepackage{tikz}`  [tikz Examples](https://texample.net/tikz/examples/all/), [tikz 中文手册-1000多页](https://www.latexstudio.net/uploads/20200703/ManualNotes315b0703.pdf)

![](https://fig-lianxh.oss-cn-shenzhen.aliyuncs.com/20210811093944.png)

--- - --

![](https://fig-lianxh.oss-cn-shenzhen.aliyuncs.com/20210819130451.png)

![](https://fig-lianxh.oss-cn-shenzhen.aliyuncs.com/20210819130729.png)

--- - --

### 习惯：可重现
> [可重复性研究：如何保证你的研究结果可重现？](https://www.lianxh.cn/news/6d3f9bbbdef36.html) 
- 文件夹：一篇文章一个文件夹；设定根路径；`readme.txt`
- 代码：`version 16`、格式、`help net`&rarr;`net set`
- 可重复 &rarr; 效率&emsp; see [论文复现-28篇](https://gitee.com/arlionn/collections/175903), &emsp; readme.txt !

![](https://fig-lianxh.oss-cn-shenzhen.aliyuncs.com/20210819132105.png)

--- - --

![](https://fig-lianxh.oss-cn-shenzhen.aliyuncs.com/20210819135707.png)

--- - --

## 2. 与老师 (同行) 们的讨论
  - 导师：导 v.s 教
    - 测试学生的潜力：`lianxh`, `ihelp`, `cnbeta`, `xtsmooth`, `ttsf`
    - 养成写作的习惯：推文、讲义 (Econometrica)
  - 一个导师的自我修炼：
    - Stata + Python ( R ) + ……
    - 用 Markdown 或 LaTeX 写讲义
    - 看文献 + 看代码

--- - --

## 3. 对 Stata 公司的一点建议

- 简单对比
  - [Stata - SSC](https://econpapers.repec.org/software/bocbocode/) (3605 items [Top 25](https://logec.repec.org/scripts/itemstat.pf?type=redif-software))
    - Github ？？ [github-Stata](https://github.com/search?l=Stata&o=desc&q=stata&s=stars&type=Repositories) (1,710 repositories)
  - [R packages](https://cran.r-project.org/web/packages/) (18038 packages) | [R github](https://github.com/cran) (Repositories 21.7k)
  - [Python - pypi.org](https://pypi.org/) (322,106 projects)
- 谁在用 Stata？
--- - --

- 建议
 - 一个易于分享、规范的社区 &rarr; packages (不同于 [FAQs](https://www.stata.com/support/faqs/), [Stata List](https://www.statalist.org/))
 - adofiles 发布：模板、规则
 - [Stata Journal](https://www.stata-journal.com/) &rarr; **Stata Tips**, **Stata Blogs** &rarr; **Stata Packages**
 - Markdown 支持：其他编辑器中的语法高亮
 - LaTeX 支持：
 - 已有：`sjlatex`, `stata.sty`, `esttab`, `outreg2`, [`texdoc`](https://journals.sagepub.com/doi/pdf/10.1177/1536867X1601600201)……
 - 可以有：`stata.sty` 收录到 TeXLive 发行版中；语法高亮
 - R 接口：`rcall`
 - Cheaper …… ^-^

--- - --

## &emsp; &emsp; &emsp; &emsp; &emsp; &emsp; &emsp; &emsp; &emsp; &emsp; &emsp; Thanks

Evainquixotic/五分钟Markdown

简介

发行版

贡献者

近期动态

Evainquixotic/五分钟Markdown .gitee-modal { width: 500px !important; }

简介

发行版

贡献者

近期动态

搜索帮助

Evainquixotic/五分钟Markdown