metadata

dataset_info:
  features:
    - name: id
      dtype: int64
    - name: source_text
      dtype: string
    - name: target_text
      dtype: string
    - name: proficiency
      dtype: int64
    - name: source_language_name
      dtype: string
    - name: source_language_code
      dtype: string
    - name: target_language_name
      dtype: string
    - name: target_language_code
      dtype: string
    - name: created_at
      dtype: string
    - name: created_by
      dtype: string
  splits:
    - name: train
      num_bytes: 5130209
      num_examples: 12346
  download_size: 2947702
  dataset_size: 5130209
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*
language:
  - ban
  - min
  - ace
  - bug
  - bjn
  - jv
  - su
  - nij
  - sas
  - mad
pretty_name: Indonesiaku Contributions
size_categories:
  - 1M<n<10M

Indonesiaku Contributions Dataset

Dataset Description

The Indonesiaku Contributions dataset contains user-submitted translations collected via the Indonesiaku platform. It aims to support language learning, translation research, and the development of multilingual models by providing a diverse set of source-target text pairs across various languages, enriched with metadata such as proficiency levels and contributor information.

This dataset is sourced from the Indonesiaku translation database, where users contribute translations as part of a collaborative effort to preserve and share linguistic knowledge. For more details on the platform, visit http://indonesiaku.com/data.

Key Features

Multilingual: Includes translations between multiple language pairs.
Crowdsourced: Contributions from a community of language learners and native speakers.
Rich Metadata: Contains proficiency ratings, language codes, and contributor usernames.

Dataset Structure

The dataset is provided as a single split (train) in CSV format with the following columns:

Column	Description	Type
`id`	Unique identifier for the translation	Integer
`source_text`	Original text in the source language	String
`target_text`	Translated text in the target language	String
`proficiency`	Contributor’s proficiency level (0-3)	Integer
`source_language_name`	Name of the source language	String
`source_language_code`	ISO code of the source language	String
`target_language_name`	Name of the target language	String
`target_language_code`	ISO code of the target language	String
`created_at`	Timestamp of translation submission	String
`created_by`	Username of the contributor (nullable)	String

Proficiency Levels:
- 0: Not Proficient
- 1: Elementary
- 2: Limited Working
- 3: Native/Bilingual

Usage

Loading the Dataset

You can load the dataset using the Hugging Face datasets library:

from datasets import load_dataset

dataset = load_dataset("williamhtan/indonesiaku-contrib")
print(dataset["train"][0])  # Access the first row

Example Applications

Machine Translation: Train or fine-tune translation models.
Language Learning: Analyze proficiency impacts on translation quality.
Linguistic Research: Study language pair distributions and translation patterns.

Data Collection

The data is collected from the Indonesiaku translation database, where users submit translations as part of an open contribution process. The dataset is periodically updated to reflect new contributions. This version (1.0.0) was exported on [insert date, e.g., February 23, 2025].

Licensing

This dataset is released under the MIT License.

Citation

If you use this dataset in your research or projects, please cite it as:

@dataset{williamhtan_indonesiaku_contrib,
  author = {William Tan and Indonesiaku Contributors},
  title = {Indonesiaku Contributions Dataset},
  year = {2025},
  publisher = {Hugging Face},
  url = {https://huggingface.co/datasets/williamhtan/indonesiaku-contrib}
}

Acknowledgments

Thanks to the Indonesiaku community for their contributions.
Built with support from Supabase for database hosting and Hugging Face for dataset hosting.

Contact

For questions or contributions, visit http://indonesiaku.com/data or contact the dataset maintainer at [insert contact info if desired].