brunoretiro/womanhealthfiap

Synthetic Dataset - Women's Health Analysis (English)

Description

Synthetic English dataset for multimodal analysis (text + audio) of medical consultations specialized in women's health.

Created for: Tech Challenge Module 4 - POSTECH FIAP

Structure

The dataset contains simulated medical consultations across different women's health categories, each with:

• Structured patient data (CSV/JSON)
• Report transcription (text)
• Synthetic audio recording (MP3) generated with OpenAI TTS

Categories

1. Normal Prenatal - Uncomplicated pregnancy consultations
2. Prenatal with Anxiety - Signs of gestational anxiety
3. Normal Postpartum - Postpartum recovery without complications
4. Postpartum Depression - Indicators of postpartum depression
5. Normal Gynecological - Routine gynecological visits
6. Gynecological with Suspected Violence - Subtle indicators of domestic violence
7. Menopause - Climacteric symptoms and consultations

Usage

from datasets import load_dataset

dataset = load_dataset("your-username/dataset-womens-health-en")

for example in dataset['train']:
    print(example['nome_completo'])
    print(example['relato_texto'])
    print(example['audio_path'])

Data Fields

Field	Type	Description
id	string	Unique identifier
nome_completo	string	Fictitious patient name
idade	int	Age in years
data_nascimento	string	Date of birth
telefone	string	Fictitious phone number
cidade	string	Fictitious city
estado	string	State
data_consulta	string	Consultation date
categoria_consulta	string	Consultation category
relato_texto	string	Report transcription
audio_filename	string	Audio file name
audio_path	string	Audio path
duracao_estimada_segundos	float	Estimated duration

Applications

• Sentiment analysis model training
• Emotional pattern detection in audio
• Multimodal analysis (text + audio)
• Early identification of mental health risks
• AI research applied to healthcare
• Domestic violence signal detection

Limitations

• All data is SYNTHETIC and FICTITIOUS
• Does not represent real patients
• Should not be used for real medical diagnoses
• Audio generated by OpenAI TTS (Text-to-Speech)
• For educational and research purposes only

License

MIT License - Free for educational use and research.

---

Created: February 2026 Version: 1.0 (English)