/Tobacco-on-Twitter

Source code for ICWSM'20 Spotlight paper.

Primary LanguageJupyter Notebook

Tobacco-on-Twitter

This repository contains source code for ICWSM 2020 Spotlight paper:

Quasi-experimental Designs for Assessing Response on Social Media to Policy Changes

Directories Explanation

  • Analysis: analysis of the first stage of our research ideas
  • bkResearchCluster: data preprocessing in the brooklyn research cluster
  • google_drive: data processing from google drive
  • juliana: primary directory, including data processing, figure and table generation, regression, qualitative checking and sentiment analysis for the data from juliana group
  • prince: data preprocessing in the server "prince"
  • reddit: everying relevant to reddit data
  • result: detailed record and explanation of the results
  • twitter: preprocessing of the Twitter API data
  • tobaocco: extraction of relevant tobacco data

Environment Requirement

Strongly Suggest using Anaconda to configure your environment. Each required package is indicate in front of the code. Simple requirement examples are

  • Python 3
    • pandas >= 0.18
    • numpy >= 1.11
    • sklearn = 0.17
  • Jupyter The experiments are written in Jupyter notebooks. Please follow the Jupyter installation instructions, and ensure you have version 4 or later: 
    $ jupyter --version
4.2.0

Citation

Tian, Y., & Chunara, R. (2020). Quasi-Experimental Designs for Assessing Response on Social Media to Policy Changes. Proceedings of the International AAAI Conference on Web and Social Media, 14(1), 671-682. Retrieved from https://aaai.org/ojs/index.php/ICWSM/article/view/7333