Virtual Workshop on Analysis of Unstructured Text Data

SSC - Data Science and Analytics Section

When

Thursday June 4th; 1:00-2:30 and 2:30-4:00

Github

All code and data

Instructors

Dave Campbell and Nathan Taback

R libraries to install

• Data Acquisition tools: gutenbergr, genius, rvest, RSelenium
• Tidyverse Tools: tidyverse, tidytext, dplyr, stringr, janitor, ggplot2
• LDA libraries: topicmodels, tm, lda
• Word2Vec library installs from github:
  • library(devtools)
  • install_github(“bmschmidt/wordVectors”)
  • library(wordVectors)

Part 1

Topics

• R Libraries for Working with Text
• Importing unstructured text into R: copy/paste, external API, webscraping, R libraries
• Regular expressions and text normalization (e.g., tokenization)
• N-Grams
• Word Vectors and Matricies

Slides and Materials

Slides - html

Slides - pdf

Slides - Rmd (xaringan), css style file

Justin Trudeau tweets

stringr cheatsheet

Part 2

Topics

Statistical Models for Unstructured Text

• Sentiment analysis models
• Topic modelling; LDA and what about inference?
• Word2Vec, what it is and what about inference
• Open problems and general discussion

Slides and Materials

LDA - md, LDA - html

Sentiment-Analysis - md, Sentiment-Analysis - html

Word2Vec - md, Word2Vec - html