Automated parsing and geocoding of Telegram news channels
Telegram has become one of the main platforms Open Source Intelligence (OSINT) researchers use to obtain primary information and breaking news. However, following a large number of channels is time-consuming, and extracting key information can be a tedious process. Moreover, the textual format of the data lacks a spatial component and thus is unsuitable for geographic analysis. TGeocoder leverages the Telegram API and OpenAI's GPT-4o LLM model to automatically parse large numbers of Telegram messages, and convert them into a machine-readable tabular format, wherein events within each message are summarised. It then extracts location data contained within the messages and finds the Lat/Lon coordinates of each event. Finally, each of these events is plotted on an interactive Folium map.
- Python 3.13
- conda: The installation script that installs the dependencies needs to use both conda and pip to fetch the required dependencies, so please use conda and create a new conda virtual environment.
- OpenAI: You will need an OpenAI API Key for the script to run.
- Telegram: A Telegram API ID, API Hash, and telephone number are also required.
- Create the conda environment. This will install all necessary package dependencies too.
conda env create -f environment.yml- Activate the conda environment created.
conda activate tgeocoder- Open the file titled
TGeocoder.ipynb - Add your Telegram and OpenAI API information
- Add the name of the Telegram channel you wish to parse, as well as the number of messages
- Run the notebook