README.md

---
license: mit
language:
- ru
tags:
- nenets
- low-resource
- parallel-corpus
- machine-translation
- nlp
- uralic
- endangered-language
task_categories:
- translation
- text-generation
---

# Nenets Language Datasets

📂 Source code on [GitHub](https://github.com/shnyakov/nenets)

**Abstract:**  
The *Nenets Language Datasets* project provides parallel and monolingual corpora for Tundra and Forest Nenets.  
The dataset supports research in machine translation, morphological analysis, and language modeling under low-resource conditions.

---

## Описание проекта

**Nenets Language Datasets** — это собрание параллельных и монолингвальных корпусов для изучения и обработки ненецкого языка (тундровый и лесной).  
Корпуса включают тексты из фольклора, газетных публикаций и социальных сетей.  
Проект направлен на создание открытой лингвистической базы для задач машинного перевода, морфологического анализа и языкового моделирования в условиях ограниченных данных.

---

## Motivation / Мотивация

The Nenets language belongs to the Samoyedic branch of the Uralic language family and is considered **low-resource and endangered**.  
This dataset aims to support:

- development of **neural Nenets–Russian MT systems**;  
- training of **language models** for low-resource languages;  
- **morphological and syntactic analysis** tasks;  
- **documentation and preservation** of Arctic indigenous linguistic heritage.

---

## Dataset Composition / Состав корпуса

### Tundra Nenets language / Тундровый ненецкий язык

| File | Description | Language(s) | Size |
|------|--------------|-------------|------|
| `parallel_corpus_tales_yangasova.csv` | Parallel corpus of Nenets fairy tales *Ненецие” лаханако”* with Russian translations (N.M. Yangasova collection, 2007). | Nenets ↔ Russian | 3,114 pairs |
| `articles_nvinder.csv` | News articles from *Няръяна Вындер* (nvinder.ru), sentence-aligned. | Nenets ↔ Russian | 286 pairs |

### Forest Nenets language / Лесной ненецкий язык

| File | Description | Language(s) | Size |
|------|--------------|-------------|------|
| `nn_yanao_posts.csv` | Posts from *Няръяна Нгэрм* VK community (vk.com/nn.yanao). | Nenets, Russian | 1,915 posts |
| `nn_yanao_posts_2.csv` | Posts from *«Няръяна Нгэрм». Пишем на лесном диалекте* VK community (vk.com/club173849556). | Nenets, Russian | 409 posts |
| `articles_yamal_expedition.csv` | Ethnographic texts from the project *Real People* (yamalexpedition.ru/nen/blog-nen). | Nenets | 42 articles |
| `articles_yamal_expedition_2.csv` | Ethnographic texts from *Хойняна тохолкова* (yamalexpedition.ru/nen/education-for-nomads-nen). | Nenets | 5 articles |

---

## Data Processing and Curation / Обработка данных

- Manual sentence alignment for all parallel sets.  
- Orthographic normalization and conversion to unified Cyrillic script.

For preprocessing scripts and alignment tools, see [GitHub repository](https://github.com/shnyakov/nenets).

---

## Use Cases / Применение

- **Machine Translation (MT)** Nenets ↔ Russian  
- **Language Modeling (LM)** for low-resource languages  
- **Morphological segmentation** and **POS tagging**  
- **Dialect identification** and **variation analysis**  
- **Digital humanities** and **sociolinguistic** research

---

## License and Attribution

The dataset is released under the **MIT License**.  
All text materials are used for **research and educational purposes** only.  
When using or redistributing this dataset, please cite the original sources where applicable.

---

## Citation

If you use this dataset, please cite:

```bibtex
@misc{pavel_shnyakov_2025,
	author       = { Pavel Shnyakov },
	title        = { nenets (Revision 653275c) },
	year         = 2025,
	url          = { https://huggingface.co/datasets/shnyakov/nenets },
	doi          = { 10.57967/hf/6728 },
	publisher    = { Hugging Face }
}