typingmonk's picture
Update Subset Categorization in README
d378212
metadata
language:
  - zh
configs:
  - config_name: 11-1
    data_files: csv/11-1.csv
  - config_name: 10-8
    data_files: csv/10-8.csv
  - config_name: 10-7
    data_files: csv/10-7.csv
  - config_name: 10-6
    data_files: csv/10-6.csv
  - config_name: 10-5
    data_files: csv/10-5.csv
dataset_info:
  - config_name: 11-1
    features:
      - name: ivod_id
        dtype: string
      - name: legislator_name
        dtype: string
      - name: meet_name
        dtype: string
      - name: meet_subjects
        dtype: string
      - name: meet_date
        dtype: string
      - name: speech_start_time
        dtype: string
      - name: speech_end_time
        dtype: string
      - name: gazette_agenda_content
        dtype: string
      - name: transcript
        dtype: string
  - config_name: 10-8
    features:
      - name: ivod_id
        dtype: string
      - name: legislator_name
        dtype: string
      - name: meet_name
        dtype: string
      - name: meet_subjects
        dtype: string
      - name: meet_date
        dtype: string
      - name: speech_start_time
        dtype: string
      - name: speech_end_time
        dtype: string
      - name: gazette_agenda_content
        dtype: string
      - name: transcript
        dtype: string
  - config_name: 10-7
    features:
      - name: ivod_id
        dtype: string
      - name: legislator_name
        dtype: string
      - name: meet_name
        dtype: string
      - name: meet_subjects
        dtype: string
      - name: meet_date
        dtype: string
      - name: speech_start_time
        dtype: string
      - name: speech_end_time
        dtype: string
      - name: gazette_agenda_content
        dtype: string
      - name: transcript
        dtype: string
  - config_name: 10-6
    features:
      - name: ivod_id
        dtype: string
      - name: legislator_name
        dtype: string
      - name: meet_name
        dtype: string
      - name: meet_subjects
        dtype: string
      - name: meet_date
        dtype: string
      - name: speech_start_time
        dtype: string
      - name: speech_end_time
        dtype: string
      - name: gazette_agenda_content
        dtype: string
      - name: transcript
        dtype: string
  - config_name: 10-5
    features:
      - name: ivod_id
        dtype: string
      - name: legislator_name
        dtype: string
      - name: meet_name
        dtype: string
      - name: meet_subjects
        dtype: string
      - name: meet_date
        dtype: string
      - name: speech_start_time
        dtype: string
      - name: speech_end_time
        dtype: string
      - name: gazette_agenda_content
        dtype: string
      - name: transcript
        dtype: string
license: apache-2.0

Taiwan Legislator Transcript

Overview

The transcripts of speech record happened in various kinds of meetings at Taiwan Legislator Yuan. The original of transcripts are compiled and published on gazettes from Taiwan Legislator Yuan.

For each segment of transcript, there are corresponding video clip on Legislative Yuan IVOD system. IVOD stands for Internet Video on Demand system.

For more detail on data origin please look at:

Subset Categorization

The naming of subset is after the rule {term}-{session} which means:

  • Term: Fixed period during which the elected legislators serve. The duration is 4 year for each term.
  • Session: From Knowledge Base of Taiwan Legislative Yuan, session means:

    Each year the Legislative Yuan shall hold two sessions of its own accord during the following Constitution regulated periods: from February to the end of May, and from September to the end of December

For each term 8 sessions is expected.

eg. 10-1, 10-2, ... , 10-8

Data Fields

Field name Description
ivod_id ID of IVOD system. The original video clip url would be https://ivod.ly.gov.tw/Play/Clip/1M/{ivod_id}
legislator_name The speaker, the name of the legislator
meet_name Name of the meeting
meet_subjects The subjects of the meeting
meet_date Date of the meeting in form YYYY-mm-dd
speech_start_time Starting timecode of the video clip segment throughout the entire meeting
speech_end_time Ending timecode of the video clip segment throughout the entire meeting
gazette_agenda_content The agenda which provided from gazette
transcript The full-text transcript

Sponsorship

The work is sponsored by 【g0v 零時小學校】繁體中文AI 開源實踐計畫

Contact

If you have any issue on the dataset. Please leave a discussion on it or contact us via:

報導者(The Reporter) data@twreporter.org

歐噴有限公司(OpenFun Ltd.) contact@openfun.tw