electricsheepafrica/nigerian_realestate_mortgage_applications

Nigeria Real Estate – Mortgage Applications

Dataset Description

Synthetic Mortgages & Financing data for Nigeria real estate sector.

Category: Mortgages & Financing
Rows: 80,000
Format: CSV, Parquet
License: MIT
Synthetic: Yes (generated using reference data from PropertyPro, Knight Frank, NBS, CBN, FMBN)

Dataset Structure

Schema

• id: string
• date: string
• city: string
• value: float
• category: string

Sample Data

| id           | date       | city          |       value | category   |
|:-------------|:-----------|:--------------|------------:|:-----------|
| REC-00067356 | 2024-01-10 | Calabar       | 3.46608e+07 | A          |
| REC-00089028 | 2024-09-26 | Lagos         | 2.83809e+07 | A          |
| REC-00786337 | 2022-08-21 | Lagos         | 3.04613e+06 | A          |
| REC-00690844 | 2024-11-25 | Lagos         | 2.3512e+07  | B          |
| REC-00836128 | 2022-12-11 | Port Harcourt | 2.84512e+07 | B          |

Data Generation Methodology

This dataset was synthetically generated using:

1. Reference Sources:

   • PropertyPro & Nigeria Property Centre - listing prices, property types, locations
   • Knight Frank & Pam Golding - market reports, prime property indices, investment analysis
   • NBS (National Bureau of Statistics) - real estate GDP contribution, construction statistics
   • CBN (Central Bank of Nigeria) - mortgage lending data, interest rates
   • FMBN (Federal Mortgage Bank) - NHF contributions, mortgage disbursements
   • State land registries - title processing times, transaction volumes

2. Domain Constraints:

   • Location pricing hierarchy (Lagos: Ikoyi > Lekki > Surulere > Ikorodu)
   • Property types (detached, semi-detached, terrace, flat, bungalow)
   • Title types (C of O, Deed of Assignment, Governor's Consent)
   • Mortgage characteristics (3% penetration, 18-25% interest, 60-70% LTV)
   • Rental yields (4-10% by location tier)

3. Quality Assurance:

   • Distribution testing (price per sqm follows lognormal within tiers)
   • Correlation validation (location-price r > 0.7, size-price r > 0.8)
   • Causal consistency (property valuation models, market dynamics)
   • Multi-scale coherence (property → neighborhood → city aggregations)
   • Ethical considerations (representative, unbiased, privacy-preserving)

See QUALITY_ASSURANCE.md in the repository for full methodology.

Use Cases

• Machine Learning: Property price prediction, location scoring, investment ROI forecasting
• Market Analysis: Price trends, supply-demand dynamics, market segmentation
• Investment: Rental yield optimization, portfolio analysis, risk assessment
• Urban Planning: Housing demand forecasting, infrastructure impact analysis
• Research: Real estate market dynamics, affordability studies, diaspora investment patterns

Limitations

• Synthetic data: While grounded in real distributions from PropertyPro/Knight Frank/NBS, individual records are not real transactions
• Simplified dynamics: Some complex interactions (e.g., negotiation, market sentiment) are simplified
• Temporal scope: Covers 2022-2025; may not reflect longer-term trends or future policy changes
• Title issues: Models title problems statistically but doesn't capture full legal complexity

Citation

If you use this dataset, please cite:

@dataset{nigeria_realestate_2025,
  title = {Nigeria Real Estate – Mortgage Applications},
  author = {Electric Sheep Africa},
  year = {2025},
  publisher = {Hugging Face},
  url = {https://huggingface.co/datasets/electricsheepafrica/nigerian_realestate_mortgage_applications}
}

Related Datasets

This dataset is part of the Nigeria Real Estate & Property Markets collection:

• https://huggingface.co/collections/electricsheepafrica/nigeria-real-estate-and-property-markets

Contact

For questions, feedback, or collaboration:

• Organization: Electric Sheep Africa
• Collection: Nigeria Real Estate & Property Markets
• Repository: https://github.com/electricsheepafrica/nigerian-datasets

Changelog

Version 1.0.0 (October 2025)

• Initial release
• 80,000 synthetic records
• Quality-assured using PropertyPro/Knight Frank/NBS/CBN reference data