Datasets:

dranreb1660
/

medimaven-qa-data

annotations_creators:
  - machine-generated
language:
  - en
license: cc-by-4.0
tags:
  - medical
  - rag
  - synthetic-qa
  - lay-symptom
pretty_name: MediMaven-QA v1.0
size_categories:
  - 100K<n<1M
dataset_info:
  config_name: qa_long
  features:
    - name: chunk_id
      dtype: string
    - name: question
      dtype: string
    - name: answer
      dtype: string
  splits:
    - name: train
      num_bytes: 52621793
      num_examples: 143280
  download_size: 26138154
  dataset_size: 52621793
configs:
  - config_name: qa_long
    data_files:
      - split: train
        path: qa_long/train-*

🩺 MediMaven-QA v1.0

MediMaven-QA is a chunk-level, citation-preserving medical question-answer corpus purpose-built for Retrieval-Augmented Generation (RAG).
It bridges everyday lay-symptom narratives with trustworthy clinical content from curated web sources.

📦 Dataset Contents

Config (`name`)	Rows	What it holds	Typical use-case
`chunks`	70 248	200-token, sentence-aware context windows with rich metadata (`id`, `url`, `title`, `section`, `source`, `n_token`, `text`)	RAG context store / retriever training
`qa_wide`	70 018	List-of-dict QA per `chunk_id` → single row may have ≥1 QA pair	Fast retrieval + generation, keeps chunk linkage
`qa_long`	143 221	Fully exploded (`chunk_id`, `question`, `answer`)	Classic supervised QA fine-tuning or eval

⚠️ Disclaimer — This corpus is for research & benchmarking only.
It is not a diagnostic tool and should not be used in clinical workflows.

🚀 Quick Load

from datasets import load_dataset

# pick one of these configs
qa_long  = load_dataset("bernard-kyei/medimaven-qa-data", "qa_long", split="train")
qa_long  = load_dataset("bernard-kyei/medimaven-qa-data", "qa_long", split="train")
# accompany with chunks to get contexts
chunks   = load_dataset("bernard-kyei/medimaven-qa-data", "kb_chunks",  split="train")

print(qa_long[0]["question"])
print(qa_long[0]["answer"])

🛠️ Generation Pipeline

Stage	Tooling	Notes
1️⃣ Crawl	Scrapy + Splash	Mayo Clinic, NHS.uk, WebMD, Cleveland Clinic (public-domain / permissive T&Cs)
2️⃣ Chunk	spaCy sentenciser	≈200 tokens / chunk; keeps heading context
3️⃣ Synthetic QA	GPT-4o-mini (`gpt-4o-mini-2024-05-preview`)	• 1 concise lay Q • 1 symptom-narrative Q → cost $40 for 143 k pairs
4️⃣ Versioning	Weights & Biases Artifacts	`kb_chunks`, `qa_wide` `qa_long`

📊 Key Stats

Metric	Value
Total context tokens	27.4 M
Avg. tokens / chunk	390
Unique host domains	4
QA pairs / chunk (mean)	2.0
% symptom-narrative Qs	51 %

🧩 Dataset Structure (Arrow schema)

click to expand

┌─────────────┬──────────────────────┐ │ chunks │ qa_wide / qa_long │ ├─────────────┼──────────────────────┤ │ id: string │ chunk_id: string │ │ url: string │ question: string │ │ title: str │ answer: string │ │ section:str │ -- qa_wide only -- │ │ source:str │ qa: list │ │ text: str │ │ │ n_token:int │ │ └─────────────┴──────────────────────┘

📜 Citation

@misc{KyeiMensah2025MediMavenQA,
  author  = {Kyei-Mensah, Bernard},
  title   = {MediMaven-QA: A Citation-Preserving Medical Q\A Dataset with Symptom Narratives},
  year    = {2025},
  url     = {https://huggingface.co/datasets/dranreb1660/medimaven-qa-data},
  note    = {Version 1.0}
}

🗒️ Changelog

Date (UTC)	Version	Highlights
2025-05-27	`v1.0`	• Sentence-aware chunking • 143 k synthetic QA pairs • Cost optimisation to $25