ESG-BERT

---

Domain Specific BERT Model for Text Mining in Sustainable Investing

How to use the code

1. Install the latest version of python Download Python

2. Clone the repository

   git clone https://github.com/shaban00/esg-bert.git

3. Change directory to the project

   cd esg-bert

4. Install python virtual environment

   python -m venv .venv

5. Activate python virtual environment

   • On Windows

     .venv\Scripts\activate

   • On Linux/MacOS

     source .venv/bin/activate

6. Install python dependencies

   pip install pandas transformers tika tqdm openpyxl

   pip install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cu118

7. Run the script

   python app.py -d 'directory1,directory2' -f 'file1,file2' -o 'filename.json'

   • You can specify either -d or -f or both
   • "-d" specifies directories separated with commas
   • "-f" specifies files separted with commas
   • "-o" specifies the JSON output filename. eg: output.json

8. Convert JSON output to CSV

   Online

   • JSON TO CSV

   Offline

   python convert.py -i "input_filename.json" -o "output_filename"

9. Import CSV into excel for analysis (optional)