This is a simple repository ('repo') meant to demonstrate the use of basic programming concepts for social science applications.
You can copy the repo to your computer using the command line, or downloading it as a zip file then unzipping it.
Go to green <> Code button and either use the Github CLI command (if you have the gh command line interface installed), or click "Download ZIP." Then unzip it, and open the .Rproj file.
We need a fairly vanilla dataset just to practice on -- ideally it will be a lot of litle files. How about the Australian Hansard?
We could try to gather it ourselves... or just google "australian parliament hansard full text github" (protip: adding 'github' to a search can sometimes turns up interesting & usefu public resources.)
In this case: https://github.com/wragge/hansard-xml
So we download a bunch of those directories. Then pop them in the ./data directory for this project.
Now we've got a little corpus of text to search through.
Just browsing the files, I saw this amusing invective:
"You are trying to lie your way out of trouble again. You are a compulsive liar.
You are the only liar who has been Prime Minister in the 17 years I have been
here."
So let's try to come up with a fuzzy search algorithm to find sentences that include an accusation of lying.
This is very rough-and-ready, but here are some potential sentences:
sentences <- c("You're a liar", "You're lying", "I've never met such a big liar
in my life", "You are trying to lie your way out of
trouble again", "You're a pathological liar", "You're a compulsive liar", "liar, liar, pants on fire!")
(Note: the pedagogical value of this exercise is the use of for loops and functions to do things to, or find things in, files... If we actually wanted to find examples of politicians accusing other politicians of dishonesty and lying, there may be other and probably better ways of doing it.)
The remainder is in the file fuzzymatch.R, which is best worked through interactively.