---
license: mit
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
dataset_info:
  features:
  - name: id
    dtype: string
  - name: url
    dtype: string
  - name: year
    dtype: int64
  - name: keywords
    list: string
  - name: source
    dtype: string
  - name: text
    dtype: string
  - name: doc_id
    dtype: string
  - name: chunk_num
    dtype: int64
  splits:
  - name: train
    num_bytes: 66844346
    num_examples: 15011
  download_size: 30128380
  dataset_size: 66844346
task_categories:
- question-answering
language:
- en
tags:
- sustainability
- policy-making
---

# Sustainability PolicyMaking Dataset 

## Description
This is a corpus dataset consisting of sustainability related documents and academic papers with a focus on UK and EU policy making. Each document is further split into chunks ranging from approximately 500-1000 words each, to help with processing and to enable QA tasks.

## Dataset Summary

- **Total Items**: 15,011
- **Average Chunk Length**: 553 words
- **Median Chunk Length**: 539 words
- **Year Range**:2000-2025
- **Sources**: ['cdn.climatepolicyradar.org', 'eur-lex.europa.eu', 'gov.uk',
       'arxiv.org', 'parliament.uk', 'civitas.eu', 'gov.uk',
       'HF-national-climate-targets-dataset', 'GOV.UK Environment Agency',
       'odyssee-mure.eu', 'assets.publishing.service.gov.uk', 'ArXiv',
       'ec.europa.eu', 'unfccc.int', 'ecolex.org', 'legislation.gov.uk',
       'gov.scot', 'register.consilium.europa.eu', 'iea.org',
       'aei.pitt.edu', 'itf-oecd.org', 'europarl.europa.eu',
       'projects.mcrit.com', 'bgs.ac.uk', 'web.archive.org',
       'climate-laws.org', 'nan', 'legislation.gov.uk', 'fca.org.uk',
       'transportenvironment.org', 'environment.ec.europa.eu',
       'climate.ec.europa.eu', 'None', 'hy4heat.info', 'gov.scot',
       'energy.ec.europa.eu', 'cbd.int', 'ofgem.gov.uk',
       'fleetnews.co.uk', 'ogauthority.co.uk', 'drivingelectric.com',
       'tfl.gov.uk', 'gov.wales', 'goultralow.com', 'pod-point.com',
       'bills.parliament.uk', 'ukpact.co.uk']


## Data Fields

Each corpora item contains the following fields:

- `id`: Unique string identifier for the chunk (str)
- `url`: Source URL of the document from which the chunk was extracted 
- `year`: Publication year (integer, nullable)
- `keywords`: List of keywords (List[str])
- `source`: Source organization or website
- `text`: Main text content 
- `doc_id`: Unique ID for each document. NB - chunks obtained from the same document will inherit the same document id. While the doc_id is unique for each document it is not unique for each chunk. 
- `chunk_num`: Chunk number within document (int)


## Authors

This dataset was developed as part of a research internship program at the University of Southampton, Department of Electronics and Computer Science, under the supervision of Shuang Ao.

**Internship Team:**
- Varun Shenoy - https://huggingface.co/varoon67
- Name - uname
- Name - uname

**Faculty Supervision:**
- [Name], PhD - uname
- [Name], PhD - uname