Intro to Text Analysis Workshop
Switch branches/tags
Nothing to show
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Type Name Latest commit message Commit time
Failed to load latest commit information.
data_novels
data_sotu
README.md
hello-python.py
ner.py
ner_sample-output.txt
sentiment.py
sentiment_sample-output.txt

README.md

README

These scripts and folders of data were used in the Intro to Text Analysis workshop led by Robin Davis at John Jay College of Criminal Justice on November 16, 2016, sponsored by the LACUNY Emerging Technologies Committee.

Slides (on Google Docs)

In this workshop, we looked at a brief history of humanities text analysis and covered the basics of using text as data. Then we explored Voyant Tools, Google Books n-gram viewer, and sentiment analysis.

Using these scripts

To use the Python scripts in this folder, you'll need to install:

You'll also need to install NLTK modules. In Python:

import nltk
nltk.download() 

A window should pop up. It might appear behind your current window. In the window, select the line beginning with book to download everything used in the book. It might take a while.


Not included in this repo: pos_list.txt and neg_list.txt, the lists of positive and negative words used in this repository. Were you at the workshop? Email Robin for the lists we used. Not at the workshop? Look online for lists of sentiment-bearing words. You'll have to run them through the PorterStemmer that comes with NLTK. Each word should be on its own line.

Note: You will probably run into some kind of errors with these scripts. That is the nature of running Python in different environments (and I haven't built in error-handling). If the errors are ignorable, then ignore them. If they won't let the script run at all, see if you can fix it by googling the error message. No luck? Sorry Charlie! I've included .txt files of sample output to see what you should have gotten.

Further work

Interested in learning more about Python-powered text analysis? Make your way through the NLTK book, freely available online and very engagingly written. It teaches you Python and natural language processing (NLP) at the same time, so you don't have to know Python to begin.

Data sources

  • data_novels/ - Project Gutenberg
  • data_sotu/ - Project Gutenberg (split up with Python); includes congressional addresses that aren't technically State of the Union addresses

Try using texts you're interested in! Use standalone .txt files (like novels from Project Gutenberg) or make a folder of .txt files.