Statistical Methods of Political Science

This repository contains materials and code for the Statistical Methods of Political Science course. The course focuses on applying statistical and computational techniques to analyze political data, covering topics such as data collection, visualization, modeling, and machine learning.

Lesson 1: Web Scraping

The first lesson introduces web scraping, a method for collecting data from websites. We use R and the libraries rvest and polite to extract structured information from web pages. The lesson covers:

• Basics of HTML structure and CSS selectors.
• Fetching data using polite requests to avoid blocking.
• Extracting news articles from government websites.
• Handling and storing scraped data in structured formats (CSV, DataFrame).

Files in this repository

• web_scraping.R – R script for extracting press releases.
• scraped_news.csv – Sample dataset of collected articles.
• README.md – Overview of the course and lesson materials.

Lesson 2: Text Analysis of Presidential Speeches

In this lesson, we explore quantitative text analysis using a corpus of U.S. presidential inaugural addresses. The goal is to apply methods from computational social science to extract insights from political speeches. We use the quanteda package in R to preprocess, analyze, and visualize textual data.

The lesson covers:

• Corpus creation and inspection using data_corpus_inaugural.
• Text preprocessing: tokenization, lowercasing, stopword removal, punctuation and number stripping.
• Construction of a Document-Feature Matrix (DFM) to represent speech content numerically.
• Frequency and diversity analysis: calculating type-token ratios across presidents.

Visualization techniques, including:

• Word clouds for overall speech content.
• Comparative word clouds grouped by presidents (e.g., Washington, Adams, Jefferson).
• Time series plots showing lexical richness over time.

(Optional) Grouped analysis by political party to assess differences in rhetoric.

This lesson demonstrates how text-as-data approaches can be used in political science to study rhetorical strategies, topic prevalence, and language evolution in political communication.

Stay tuned for future lessons on data analysis, regression models, and machine learning in political science! 🚀