Datasets:

electricsheepafrica
/

africa-synth-mental-health-psychosis-dataset-seychelles

language:
  - en
license: cc-by-4.0
tags:
  - synthetic
  - mental-health
  - africa
  - medical
  - public-health
pretty_name: Africa Psychosis Dataset
data_type: synthetic

⚠️ Synthetic dataset — Parameterized from published SSA literature, not real observations. Not suitable for empirical analysis or policy inference.

Dataset Card for Africa Psychosis Dataset

Dataset Description

Homepage: https://huggingface.co/electricsheepafrica
Repository: https://huggingface.co/electricsheepafrica/Africa-psychosis-dataset
Point of Contact: Electric Sheep Africa
License: CC-BY-4.0

Dataset Summary

The Africa Psychosis Dataset is a synthetic dataset generated by Electric Sheep Africa to model mental health presentations in African contexts. It contains 100,000 rows and 13 columns. The generation is grounded in peer-reviewed literature from Sub-Saharan Africa, capturing region-specific prevalence rates, risk factors (e.g., poverty, conflict), and cultural presentations of distress (e.g., somatic symptoms, "thinking too much").

This dataset is designed for research, educational, and methodological testing purposes. It respects the statistical properties of real-world data without exposing any Private Identifiable Information (PII), as every record is entirely synthetic.

Supported Tasks and Leaderboards

Epidemiological Modeling: Simulating disease burden across different African regions.
Health Services Research: Analyzing the gap between need and treatment access (e.g., traditional healer usage).
Risk Factor Analysis: Logistic regression testing on variables like poverty index and conflict exposure.

Languages

English (Data labels). The cultural context covers West, East, North, Southern, and Central Africa.

Dataset Structure

Data Instances

A sample instance from the dataset:

[
  {
    "region":"East Africa",
    "sex":"Male",
    "age":29.8313536542,
    "urban_rural":"Urban",
    "poverty_index":0.3536766572,
    "education_level":"Primary",
    "employment_type":"Informal",
    "household_size":6.2030524535,
    "has_psychosis":false,
    "symptom_type":null,
    "attribution_model":"Spiritual_Ancestral",
    "cannabis_use":false,
    "family_history":false
  }
]

Data Fields

Field Name	Type	Description
`region`	object	Synthetic variable based on literature distributions.
`sex`	object	Synthetic variable based on literature distributions.
`age`	float64	Synthetic variable based on literature distributions.
`urban_rural`	object	Synthetic variable based on literature distributions.
`poverty_index`	float64	Synthetic variable based on literature distributions.
`education_level`	object	Synthetic variable based on literature distributions.
`employment_type`	object	Synthetic variable based on literature distributions.
`household_size`	float64	Synthetic variable based on literature distributions.
`has_psychosis`	bool	Synthetic variable based on literature distributions.
`symptom_type`	object	Synthetic variable based on literature distributions.
`attribution_model`	object	Synthetic variable based on literature distributions.
`cannabis_use`	bool	Synthetic variable based on literature distributions.
`family_history`	bool	Synthetic variable based on literature distributions.

Data Splits

Train/Val/Test: Not applicable (Single synthetic file). All 100,000 rows are available for splitting as needed.

Dataset Creation

Curation Rationale

There is a scarcity of high-quality, open-access mental health data from African regions. This dataset aims to fill that gap by synthesizing available knowledge from systematic reviews and small-scale studies into a coherent, usable format for data scientists.

Source Data

Initial Data Collection and Normalization

This data was generated using a probabilistic pipeline keying off parameter files derived from the literature:

Ekhator, C. N., et al. (2024). Mental Health Awareness in Rural vs. Urban Areas. CARI Journals.
- Parameter: Urban/Rural Mental Health Access (Qualitative Differences)
- Region: General Africa
Patel, V., et al. (2018). The Lancet Commission on global mental health and sustainable development. The Lancet.
- Parameter: Treatment Gap & Healer Usage (>85% Treatment Gap)
- Region: Global/Africa
Mbwayo, A. W., et al. (2013). Mental health in schools in Nairobi, Kenya. Journal of Child & Adolescent Mental Health.
- Parameter: Psychotic Like Experiences (Youth) (~10-20%)
- Region: Kenya (East Africa)
Burns, J. K., et al. (2014). Psychosis in Africa: A systematic review. Schizophrenia Research.
- Parameter: Psychosis Prevalence (1% - 4.4%)
- Region: Sub-Saharan Africa
Abbo, C. (2011). Profiles and outcome of traditional healing practices for severe mental illnesses in two districts of Eastern Uganda. Global Health Action.
- Parameter: Traditional Healer Usage (High (~60% first contact))
- Region: Uganda (East Africa)

Personal and Sensitive Information

This dataset contains fake medical info. No real individuals are represented. However, it models sensitive topics (Depression, Psychosis, HIV status) and should be handled with the ethical care appropriate for health data to avoid stigmatization in downstream analyses.

Considerations for Using the Data

Social Impact of Dataset

Positive: Enables tool development for under-resourced regions.
Negative: If treated as "real" ground truth without validation, it could lead to incorrect policy assumptions. Always validate synthetic findings against local clinical data.

Discussion of Biases

Literature Bias: The data reflects the biases of the underlying studies (e.g., more data from South Africa/Nigeria/Kenya than other regions).
Simplification: Complex interactions (e.g., genetic vs environmental) are simplified into probabilistic dependencies.

Additional Information

Licensing Information

Citation Information

Please cite the underlying sources listed above when using this data for research context.