CoolFace
Datasetpublic

BrainHealthAI/MedQADataDarijaSaad

Health QA Darija — Medical QA in Moroccan Arabic (الدارجة المغربية) Dataset Description A curated dataset of 8,129 medical question-answer pairs in Moroccan Darija (الدارجة المغربية). Each entry contains a patient scenario, a focused medical question, and a doctor's response — all in authentic Darija. Enriched with named medical entities (symptoms, diseases, medications, tests). 🇲🇦 First large-scale medical QA dataset in Moroccan Darija — addressing the… See the full description on the dataset page: https://huggingface.co/datasets/BrainHealthAI/MedQADataDarijaSaad.

sourceHugging Faceapache-2.0updated 5mo agoView on Hugging Face
0likes24downloads
Dataset Card

Health QA Darija — Medical QA in Moroccan Arabic (الدارجة المغربية)

Dataset Description

A curated dataset of 8,129 medical question-answer pairs in Moroccan Darija (الدارجة المغربية). Each entry contains a patient scenario, a focused medical question, and a doctor's response — all in authentic Darija. Enriched with named medical entities (symptoms, diseases, medications, tests).

🇲🇦 First large-scale medical QA dataset in Moroccan Darija — addressing the critical gap in low-resource medical NLP for North African Arabic dialects.

Key Features

  • 8,129 entries across 27 medical specialties
  • Moroccan Darija language — written as spoken by patients
  • Structured schema: question + context_question + answer
  • Rich entity annotations: symptoms, diseases, medications, medical fields
  • Urgency classification: High / Moyen / Faible
  • Quality score: 93.8/100

Dataset Structure

Schema

FieldTypeDescription
questionstringالمريض كيسأل سؤال طبي بالدارجة
context_questionstringالسياق الطبي ديال المريض
answerstringالجواب ديال الطبيب بالدارجة
languagestringDarija
urgencystringHigh / Moyen / Faible
specialitystringالتخصص الطبي
article_titlestringعنوان الموضوع الطبي
entities_*list[str]الكيانات الطبية المستخرجة

Example

json
{
  "question": "عندي حساسية من القصاصات، شنو ندير؟",
  "context_question": "كنعاني من حساسية وبغيت نعرف شنو ندير.",
  "answer": "خاصك تجنب القصاصات وتستاشر مع طبيب مختص.",
  "urgency": "Moyen",
  "speciality": "طب الحساسية والمناعة",
  "entities_sympt": ["حساسية"]
}

Specialty Distribution

التخصصالعددالنسبة
طب الغدد والسكري85910.6%
طب الأمراض الجلدية7519.2%
الطب النفسي4755.8%
طب العيون4615.7%
الطب الباطني4495.5%
طب الأطفال4495.5%
طب القلب والشرايين4375.4%
طب الأسنان4225.2%
طب الأعصاب4175.1%
طب التخدير4155.1%
طب التغذية3844.7%
طب العظام والمفاصل3844.7%
طب الأذن والأنف والحنجرة3614.4%
طب الأورام3544.4%
طب أمراض الدم3514.3%
الطب العام3404.2%
طب النساء والتوليد3224.0%
طب الأمراض الصدرية1461.8%
طب الحساسية والمناعة1291.6%
طب الأمراض المعدية851.0%
طب الجهاز الهضمي470.6%
الجراحة العامة450.6%
طب المسالك البولية والتناسلية350.4%
الطب الوراثي60.1%
الطب البيطري30.0%
علم السموم10.0%
العلاج الطبيعي10.0%

Urgency Distribution

الاستعجالالعددالنسبة
Moyen (متوسط)541666.6%
Faible (منخفض)163220.1%
High (مرتفع)108113.3%

Quality Metrics

MetricValue
Overall Quality Score93.8/100
Field Completeness100%
Entity Coverage99.7%
Vocabulary Size21,447 unique words
Specialty Balance0.90
Identical Q/CQ0

Intended Use

  • Medical QA fine-tuning for Darija language models
  • Low-resource NLP research for Moroccan Arabic
  • Medical chatbot development for Moroccan patients
  • Cross-lingual medical NLP (paired with HealthQAEnglish)

Limitations

  • Synthetic Darija text (GPT-generated from English medical QA)
  • Smaller dataset compared to English counterpart
  • Some specialties underrepresented
  • Not for direct medical advice — research only

Related Datasets

Citation

bibtex
@dataset{health_qa_darija_2026,
  title={Health QA Darija: Medical Question Answering in Moroccan Arabic},
  author={Saad Karzabi},
  year={2026},
  publisher={Hugging Face},
  url={https://huggingface.co/datasets/BrainHealthAI/MedQADataDarijaSaad}
}

License

Apache 2.0