Omar-youssef/English-Egyptian-Arabic-Translation-Pairs

Bilingual Egyptian Dataset

Dataset Description

This dataset contains bilingual text pairs in English and Egyptian Arabic / Modern Standard Arabic. Each entry provides parallel translations covering a wide range of everyday and specialized domains, making it valuable for translation models, multilingual NLP research, and general Arabic language applications.

Key Features

• Languages: English ↔ Egyptian Arabic (العامية المصرية) & Arabic (العربية)
• Domain: Multi-domain (general, social, cultural, professional, and more)
• Format: Parallel text pairs with language variant labeling
• Content: Diverse topics spanning everyday conversation to specialized domains
• Quality: Professional-level translations maintaining context and meaning
• Size: 8,630 parallel pairs

Dataset Structure

Data Fields

• English: Content written in English
• Translation: Corresponding translation in Egyptian Arabic or Arabic dialect
• language: Language variant of the translation — either Egyptian Arabic or Arabic

Example

{
  "English": "What are the specific steps to inquire about the Integrated Services Card via the Ministry of Social Solidarity's website using the National ID?",
  "Translation": "إيه هي الخطوات المحددة للاستعلام عن كارت الخدمات المتكاملة من خلال الـ website بتاع وزارة التضامن الاجتماعي بالرقم القومي؟",
  "language": "Egyptian Arabic"
}

{
  "English": "The acquisition of knowledge is a lifelong journey that empowers individuals to critical thinking, problem-solving, and innovation, ultimately serving as the bedrock for a nation's progress and its ability to adapt to global challenges like Climate Change or Economic Shifts, fostering sustainable development.",
  "Translation": "اكتساب المعرفة هو رحلة مدى الحياة تمكّن الأفراد من التفكير النقدي وحل المشكلات والابتكار، مما يخدم في نهاية المطاف كأساس لتقدم الأمة وقدرتها على التكيف مع التحديات العالمية مثل Climate Change أو Economic Shifts، مما يعزز التنمية المستدامة.",
  "language": "Arabic"
}

Usage

Loading the Dataset

from datasets import load_dataset

# Load the dataset
dataset = load_dataset("Omar-youssef/English-Egyptian-Arabic-Translation-Pairs")

# Access the training split
train_data = dataset['train']

print(f"Total examples: {len(train_data)}")  # 8,630

# Example of accessing parallel text pairs
for example in train_data.select(range(3)):
    print(f"English: {example['English']}")
    print(f"Translation: {example['Translation']}")
    print(f"Language: {example['language']}")
    print("-" * 50)

Data Quality Features

• Natural Translation: Maintains Egyptian Arabic dialect authenticity where applicable
• Dialect Coverage: Includes both Egyptian colloquial Arabic and Modern Standard Arabic
• Multi-Domain: Covers a broad range of topics and domains
• Cultural Context: Adapts content to Egyptian and Arab cultural context
• Consistent Terminology: Uses standardized terminology across domains
• Comprehensive Coverage: Spans multiple everyday and specialized topics

Citation

If you use this dataset in your research, please cite:

@dataset{omar_youssef_bilingual_egyptian,
  author = {Omar Youssef},
  title = {Bilingual Egyptian Arabic Dataset},
  year = {2026},
  publisher = {Hugging Face},
  url = {https://huggingface.co/datasets/Omar-youssef/English-Egyptian-Arabic-Translation-Pairs},
  note = {English–Egyptian Arabic / Arabic parallel corpus, multi-domain, 8630 pairs}
}

License

This dataset is released under the Apache 2.0 license, allowing for both research and commercial use.