Network Analysis of German COVID-19 Related Discussions on Telegram

Authors: Valentin Peter, Ramona Kühn, Jelena Mitrović, Michael Granitzer and Hannah Schmid-Petri.

Setup project on Mac OS

To install the required dependencies and packages run:

1. Run Setup script sh setup.sh
2. Activate virtual enviroment source televenv/bin/activate
3. Start jupyter notebook jupyter-notebook

Scraper

The Scraper uses a technique called discriminative snowball sampling to sample Telegram channels given a seed. It can be accessed from within the repository as such:

# This is an example how the code could be used
from scraper.scraper import Scraper
from scraper.client import Client

Client(api_id=123456,
       api_hash="***********",
       phone="+49***********")

scraper = Scraper(num_messages=None,
                  step_size=200,
                  maximum_iterations=1)

scraper.scrape(["some_channel", "some_other_channel"])

Data analysis

The analysis of over 50 million messages is conducted in the two notebooks [Data Preprocessing](Data Preprocessing.ipynb) and final_analysis. The Data Preprocessing notebook needs to be executed first, the analysis notebook second. Due to privacy reasons the data wont be published. When using the same seed as described in the thesis on the scraper it is possible to reproduce the data set. The code for reproducing the data is provided in the file example.py.

The original execution ist still maintained in final_analysis, therefore code and results can be inspected simultaneously.