metadata

language_creators:
  - crowdsourced
language:
  - en
license: cc-by-nc-sa-4.0
size_categories:
  - 10M<n<100M
task_categories:
  - question-answering
pretty_name: Twitter Customer Support Dialogue in Kaggle
dataset_info:
  features:
    - name: created_at
      dtype: string
    - name: text
      dtype: string
    - name: dialogue_id
      dtype: int64
    - name: company_author
      dtype: string
  splits:
    - name: train
      num_bytes: 362719425
      num_examples: 794335
  download_size: 203369061
  dataset_size: 362719425
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*

Customer Support on Twitter in Kaggle

Multi-turn dialogue dataset created from Customer Support on Twitter dataset in Kaggle. Conversation is grouped in a row, each Tweet is annotated by "User_{author_id}: " or "Support: " prefix.