社會科學領域的資料科學學群規劃

目的

以資料為核心，規劃出能培養以下3面向能力的學群課程：

• 資料分析

• 調查方法

• 應用程式開發

課程類型

統計類：統計、計量

程式類：R，Python，C++

實證類：經濟、社會實證相關

生活應用類：App、網頁設計、資料視覺化

社會學系

經濟學系

挑戰：連結與延伸

• 如何與不同老師的專長結合（不只是在資料實證領域，包含理論研究）。

• 多樣的程式語言（R,Python,STATA），如何有一個較為統一的程式寫作環境。

核心程式寫作環境

寫作環境是一切的連結點。

• RStudio為寫作環境

• 採用R Markdownd格式

What is R Markdown

• Markdown

  A lightweight markup language that you can use to add formatting elements to plaintext text documents.

  • Markdown guide

  • Online editor

• R Markdown

  Documents written in markdown (an easy-to-write plain text format) with chunks of embedded R code.

寫作

R Markdown

• 容易上手

  • 已懂LaTex者：上手容易（內文寫法雷同）

  • 不懂LaTex者：比LaTex易懂

• 程式、圖、表穿插容易

• 可生成多種檔案格式：Word, PDF, html等

• 網路分享協作容易

• RStudio可單純當做markdown寫成及文件生成環境，無需懂R。

寫作：範例

---
title: "Untitled"
short: "A shorter title"
journal: "AER" # AER, AEJ, PP, JEL
month: "6"
year: "2018"
vol: 1
issue: 1
jel:
  - A10
  - A11
keywords:
  - first keyword
  - second keyword
author:
  - name: Alice Anonymous
    firstname: Alice
    surname: Anonymous
    email: alice@example.com
    affiliation: Some Institute of Technology
  - name: Bob Security
    firstname: Bob
    surname: Security
    email: bob@example.com
    affiliation: Another University
acknowledgements: |
  Acknowledgements
abstract: |
  Abstract goes here
output: rticles::aea_article
---

American Economic Review Pointers:

\begin{itemize}
\item Do not use an "Introduction" heading. Begin your introductory material
before the first section heading.

\item Avoid style markup (except sparingly for emphasis).

\item Avoid using explicit vertical or horizontal space.

\item Captions are short and go below figures but above tables.

\item The tablenotes or figurenotes environments may be used below tables
or figures, respectively, as demonstrated below.

\item If you have difficulties with the mathtime package, adjust the package
options appropriately for your platform. If you can't get it to work, just
remove the package or see our technical support document online (please
refer to the author instructions).

\item If you are using an appendix, it goes last, after the bibliography.
Use regular section headings to make the appendix headings.

\item If you are not using an appendix, you may delete the appendix command
and sample appendix section heading.

\item Either the natbib package or the harvard package may be used with bibtex.
To include one of these packages, uncomment the appropriate usepackage command
above. Note: you can't use both packages at once or compile-time errors will result.

\end{itemize}

\section{First Section in Body}

Sample figure:

\begin{figure}
Figure here.

\caption{Caption for figure below.}
\begin{figurenotes}
Figure notes without optional leadin.
\end{figurenotes}
\begin{figurenotes}[Source]
Figure notes with optional leadin (Source, in this case).
\end{figurenotes}
\end{figure}

Sample table:

\begin{table}
\caption{Caption for table above.}

\begin{tabular}{lll}
& Heading 1 & Heading 2 \\
Row 1 & 1 & 2 \\
Row 2 & 3 & 4%
\end{tabular}
\begin{tablenotes}
Table notes environment without optional leadin.
\end{tablenotes}
\begin{tablenotes}[Source]
Table notes environment with optional leadin (Source, in this case).
\end{tablenotes}
\end{table}

References here (manual or bibTeX). If you are using bibTeX, add your bib file
name in place of BibFile in the bibliography command.
% Remove or comment out the next two lines if you are not using bibtex.
\bibliographystyle{aea}
\bibliography{references}

% The appendix command is issued once, prior to all appendices, if any.
\appendix

\section{Mathematical Appendix}

投影片製作

• ioslides, Slidy, Beamer, Reveal.js

  • RStudio gallery

  • Reveal.js範例

網路發佈

• bookdown.org

Git and Github

Git:

A revision control system, a tool to manage your source code history.

GitHub:

A hosting service for Git repositories.

Git and RStudio

說明

版本追踪

每工作一段落可設版本記錄點(commit)，之後可輕易回復之前任何版本。

網路協作

使用Github來完成

• RStudio：本機版本追踪

• Github：雲端備份與協作

網路協作：範例

Repo network

設定: Video 1

設定: Video 2

RStudio的多語言連結

透過不同套件可在R Markdown裡穿插使用多種程式語言。

• Stata: RStata

• Python: reticulate (說明)

• C++: Rcpp

• Julia: JuliaCall (說明)

Julia

a high-level, high-performance dynamic programming language for numerical computing. It provides a sophisticated compiler, distributed parallel execution, numerical accuracy, and an extensive mathematical function library.

• 數值分析工具

• 程式執行速度與C++不相上下

• 使用Maxima.jl套件可進行與Mathematica相同的數學符號運算。

Julia: RStudio使用

透過knitr

knitr::opts_chunk$set(echo = TRUE)
knitr::knit_engines$set(julia = JuliaCall::eng_juliacall)

a=sqrt(5)

## 2.23606797749979

print(a)

經濟理論模型應用

• Lectures in Quantitative Economics, Thomas J. Sargent and John Stachurski.

• AN INTRODUCTION TO STRUCTURAL ECONOMETRICS IN JULIA

“The contrast is staggering: the optimization that didn’t converge after 24 hours in Python converged after only 15 minutes in Julia while Python was still running on the same processor.” – Bradley Setzler.

總結

統計、計量：R

機器學習：Python

經濟數學、追求數學運算速度：Julia

寫作介面：RStudio/R Markdown

同時使用Github進行版本追蹤與協作