ENTRANT: A Large Financial Dataset for Table Understanding

Extract and clean tables from financial xlsx files from EDGAR and convert them to JSON with bi-tree positional information and metadata.

Related dataset:

Zavitsanos, E., Mavroeidis, D., Spyropoulou, E., Fergadiotis, M., & Paliouras, G. (2024). ENTRANT: A Large Financial Dataset for Table Understanding [Data set]. Zenodo. https://doi.org/10.5281/zenodo.10667088

Please cite the related paper as follows: Zavitsanos, E., Mavroeidis, D., Spyropoulou, E. et al. ENTRANT: A Large Financial Dataset for Table Understanding. Sci Data 11, 876 (2024). https://doi.org/10.1038/s41597-024-03605-5

Table of Contents

• Install
• Usage
• Data
• Tests
• Contributing
• License
• Citation

Install

• Before starting, ideally, it's recommended to switch to a virtual environment first via conda or virtualenv.
• Install dependencies via pip install -r requirements.txt

Usage

For table extraction from EDGAR:

• Place the xls files in a directory named data in the project's root.
• Create a directory named output to store the results.
• Run extract_tables_multiprocess.py.

For downloading excel reports

• See fetch_reports.py
• Pay attention to fair usage of EDGAR

Data

• Data is hosted at Zenodo: https://zenodo.org/records/10667088

Tests

Use pytest to run the unit tests.

Contributing

See the contributing file!

License

The project is licensed under Creative Commons Attribution 4 license.

Citation

@article{entrant2024,
  title={ENTRANT: A Large Financial Dataset For Table Understanding},
  author={Zavitsanos, Elias and Mavroeidis, Dimitris and Spyropoulou, Eirini and Fergadiotis, Manos and Paliouras Georgios},
  journal={Nature Scientific Data},
  pages={876},
  volume = {11},
  year={2024},
  doi = {https://doi.org/10.1038/s41597-024-03605-5}
}