---
license: cc-by-4.0
task_categories:
- fill-mask
- text-classification
language:
- en
tags:
- business
- company website
- industry classification
pretty_name: CompanyWeb
size_categories:
- 1M<n<10M
task_ids:
- masked-language-modeling
---

# Dataset Card for "CompanyWeb"

### Dataset Summary

The dataset contains textual content extracted from 1,788,413 company web pages of 393,542 companies. The companies included in the dataset are small, medium and large international enterprises including publicly listed companies. Additional company information is provided in form of the corresponding Standard Industry Classification (SIC) label `sic4`. 
The text includes all textual information contained on the website with a timeline ranging from 2014 to 2021. The search includes all subsequent pages with links from the homepage containing the company domain name. 
We filter the resulting textual data to only include English text utilizing the FastText language detection API [(Joulin et al., 2016)](https://aclanthology.org/E17-2068/). 

### Languages

- en

## Dataset Structure

### Data Instances

- **#Instances:** 1789413
- **#Companies:** 393542
- **#Timeline:** 2014-2021

### Data Fields

- `id`: instance identifier `(string)`
- `cid`: company identifier `(string)`
- `text`: website text `(string)`
- `sic4`: 4-digit SIC `(string)`

### Citation Information

```bibtex
@article{BORCHERT2024,
  title = {Industry-sensitive language modeling for business},
  journal = {European Journal of Operational Research},
  year = {2024},
  issn = {0377-2217},
  doi = {https://doi.org/10.1016/j.ejor.2024.01.023},
  url = {https://www.sciencedirect.com/science/article/pii/S0377221724000444},
  author = {Philipp Borchert and Kristof Coussement and Jochen {De Weerdt} and Arno {De Caigny}},
}
```