ayookuns/abuja-housing-prices-v1

Abuja Housing Prices — Groundwork Data v1

Dataset Description

This is the first public, structured dataset of Abuja residential property prices — covering both rental and sale listings across 16 areas of Nigeria's Federal Capital Territory.

Collected by Groundwork Data, a Nigerian civic data project building the structured, domain-specific datasets that Nigerian ML engineers, researchers, and startups actually need.

Nigeria has a severe structured data gap. If you want to build something that touches how Nigerians live, rent, buy, or move — you will hit a wall. Every time. This project is fixing that.

• Homepage: https://www.kaggle.com/datasets/ayoolaokunlola/abuja-housing-prices-groundwork-data-v1
• License: CC BY 4.0
• Version: v1 (April 2026)

---

Dataset Summary

Listings	481
Rental listings	250
Sale listings	231
Areas covered	16
Property types	7
Price range	₦800,000 – ₦5,000,000,000
Source	PropertyPro.ng
Scraped	April 2026

---

Data Fields

Column	Type	Description
title	string	Listing headline as displayed on PropertyPro.ng
price_ngn	int	Listed price in Nigerian Naira. Annual rent for rental listings, asking price for sale listings
listing_type	string	"rent" or "sale"
area	string	Broad Abuja area/district. 16 unique values
neighbourhood	string	Sub-area within district. 17 unique values
address	string	Street-level address string as listed
bedrooms	int	Number of bedrooms. Range: 1–10
bathrooms	float	Number of bathrooms. Range: 1–10. ~2% missing
toilets	float	Number of toilets. Range: 1–10. ~18% missing
property_type	string	Flat/Apartment, Detached Duplex, Terraced Duplex, Semi-Detached Duplex, Bungalow, House, Mansion
description	string	Note: all entries contain a generic PropertyPro safety disclaimer — not property-specific
pid	string	Unique PropertyPro listing ID. Use for deduplication
date_added	string	Date listing was first published. Format: DD-Mon-YY
last_updated	string	Date listing was last updated. Format: DD-Mon-YY. 3 missing
source_url	string	Full URL of original PropertyPro.ng listing

---

Key Findings

• Median annual rent across Abuja: ~₦8M. Maitama and Asokoro exceed ₦50M median.
• Median sale price: ~₦250M. Asokoro (₦1.1B) and Wuse 2 (₦950M) are the most expensive markets.
• A 3-bedroom property in Asokoro has a median sale price 13x higher than the same spec in Kubwa.
• Bedrooms and area location are the dominant price predictors.
• Flat/Apartment is the most common property type at 36% of listings.
• Baseline Random Forest models: CV R² = 0.79 (rent), 0.71 (sale).

---

Known Limitations

• No property size (sqm): Not consistently listed on PropertyPro. Most significant missing feature for modelling.
• No geo-coordinates: Area names only, no lat/lng.
• Single source: v1 covers PropertyPro.ng only. NigeriaPropertyCentre added in v2.
• Snapshot: Single scrape from April 2026. Not a time series.
• Description field: Generic disclaimer text only — not useful as a feature.

---

Usage

import pandas as pd
df = pd.read_csv("groundwork_abuja_housing_v1.csv")

# Separate rent and sale
rent = df[df.listing_type == "rent"]
sale = df[df.listing_type == "sale"]

# Median price by area
print(rent.groupby("area")["price_ngn"].median().sort_values(ascending=False))

Full baseline model notebook available on Kaggle: https://www.kaggle.com/datasets/ayoolaokunlola/abuja-housing-prices-groundwork-data-v1/code

---

Citation

@dataset{groundworkdata2026abuja,
  author    = {Groundwork Data},
  title     = {Abuja Housing Prices v1},
  year      = {2026},
  publisher = {Kaggle},
  url       = {https://www.kaggle.com/datasets/ayoolaokunlola/abuja-housing-prices-groundwork-data-v1}
}

---

About Groundwork Data

Groundwork Data is building Nigerian civic data infrastructure — scraping, cleaning, structuring, and publishing the domain-specific datasets that local engineers and researchers need to build things that actually work for Nigerian problems.

Follow on Kaggle and Hugging Face: Groundwork Data

Coming next: Lagos Housing Prices v1, Nigerian Schools by LGA.