maor63/juan-vocabulary-based-method-for-quantify-controversy

Vocabulary-based-Method-for-Quantify-Controversy

Python and R code to identify automatically controversy on Twitter through text

Requirements

• R version 3.6.0

• R libraries

  • igraph
  • rtweet
  • tidyverse
  • readr
  • rgexf

• Python 3

• Python libraries

  • networkx
  • numpy
  • scipy

  Usage

• File "crear_grafo.R" download data by a hashtag or key word, creates the retweet graph and save enviroment

• Bash file "calcular" recieves as parameter the name of the saved enviroment and do the following steps:

  • Create training and test files for Fasttext
  • Train the NLP model
  • Predict tags
  • Measure controversy

-It is important to give permission to create files inside the folder and to give execution permission to file calcular if you want to use it

• Example: ./calcular bigil

  Datasets

Tweet's ids used in the paper are in dataset folder grouped by discussion. New datasets collected by crear_grafo.R will store in same folder.

In file datasets/Discussion descriptions.pdf there are detail descriptions about each discussion