metadata

dataset_info:
  features:
    - name: section
      dtype: string
    - name: question
      dtype: string
    - name: answer
      dtype: string
    - name: hash
      dtype: string
    - name: source
      dtype: string
  splits:
    - name: train
      num_bytes: 5018528
      num_examples: 8775
  download_size: 1947011
  dataset_size: 5018528
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*
language:
  - en
pretty_name: Data management plans questions and answers

Info

Dataset contains cleaned questions and answers from public dmps on https://dmponline.dcc.ac.uk/

Data modification

QAs have been parsed from pdfs
Non-English dmps have been removed
Auto-generated sentences have been removed (dmponline allows some answers to be answered by selecting a predefined answer from a list. These predefined answers have been removed, only user written answers were kept.)
Answers with occurrence count greater than 8 have been removed. (This has filtered out answers such as 'No', 'Yes', 'Not applicable.', ...)
Answers with length less or equal to 3 were removed