electricsheepafrica/africa-cod-ps-global

OCHA Global Subnational Population Statistics

Publisher: HDX · Source: HDX · License: cc-by · Updated: 2026-04-03

---

Abstract

Latest COD population statistics compiled at the admin level. The CSV files contain subnational p-codes, their corresponding administrative names, source organization, and reference dates where available. These are constructed from individual country level population files, which can be found using this search on HDX.

Each row in this dataset represents country-level aggregates. Data was last updated on HDX on 2026-04-03. Geographic scope: AFG, ALB, AGO, AIA, ATG, ARG, ARM, ABW, and 135 others.

Curated into ML-ready Parquet format by Electric Sheep Africa.

---

Dataset Characteristics

Domain	Demographics and population
Unit of observation	Country-level aggregates
Rows (total)	6,722
Columns	13 (4 numeric, 9 categorical, 0 datetime)
Train split	5,377 rows
Test split	1,344 rows
Geographic scope	AFG, ALB, AGO, AIA, ATG, ARG, ARM, ABW, and 135 others
Publisher	HDX
HDX last updated	2026-04-03

---

Variables

Geographic — iso3 (BDI, TJK, SAU), country (Burundi, Tajikistan, Saudi Arabia), population_group (T_TL, F_TL, M_TL), age_max (range 0.0–99.0), population (range 0.0–269603430.0) and 1 others.

Demographic — gender (all, m, f), age_range (all, 20-24, 25-29), age_min (range 0.0–100.0).

Identifier / Metadata — source (UNFPA, Eurostat, Institut national de la statistique et des études économiques), esa_source (HDX), esa_processed (2026-04-05).

Other — contributor (UNFPA, OCHA Latin America and the Caribbean (ROLAC), OCHA Regional Office for Asia and the Pacific (ROAP)).

---

Quick Start

from datasets import load_dataset

ds    = load_dataset("electricsheepafrica/africa-cod-ps-global")
train = ds["train"].to_pandas()
test  = ds["test"].to_pandas()

print(train.shape)
train.head()

---

Schema

Column	Type	Null %	Range / Sample Values
iso3	object	0.0%	BDI, TJK, SAU
country	object	0.0%	Burundi, Tajikistan, Saudi Arabia
population_group	object	0.0%	T_TL, F_TL, M_TL
gender	object	0.0%	all, m, f
age_range	object	0.0%	all, 20-24, 25-29
age_min	float64	6.1%	0.0 – 100.0 (mean 40.1938)
age_max	float64	11.6%	0.0 – 99.0 (mean 41.7384)
population	int64	0.0%	0.0 – 269603430.0 (mean 1834614.4963)
reference_year	int64	0.0%	2001.0 – 2025.0 (mean 2020.8509)
source	object	0.0%	UNFPA, Eurostat, Institut national de la statistique et des études économiques
contributor	object	0.0%	UNFPA, OCHA Latin America and the Caribbean (ROLAC), OCHA Regional Office for Asia and the Pacific (ROAP)
esa_source	object	0.0%	HDX
esa_processed	object	0.0%	2026-04-05

---

Numeric Summary

Column	Min	Max	Mean	Median
age_min	0.0	100.0	40.1938	40.0
age_max	0.0	99.0	41.7384	39.0
population	0.0	269603430.0	1834614.4963	159803.5
reference_year	2001.0	2025.0	2020.8509	2022.0

---

Curation

Raw data was downloaded from HDX via the CKAN API and converted to Parquet. Column names were lowercased and standardised to snake_case. Common missing-value markers (N/A, null, none, -, unknown, no data, #N/A) were unified to NaN. 8 column(s) with >80% missing values were removed: adm1_pcode, adm1_name, adm2_pcode, adm2_name, adm3_pcode, adm3_name.... The dataset was split 80/20 into train and test partitions using a fixed random seed (42) and saved as Snappy-compressed Parquet.

---

Limitations

• Data originates from HDX and has not been independently validated by ESA.
• Automated cleaning cannot correct for misreported values, definitional inconsistencies, or sampling bias in the original collection.
• This dataset spans 143 countries; geographic and methodological inconsistencies across national boundaries may affect cross-country comparability.
• Refer to the original HDX dataset page for the publisher's own methodology notes and caveats.

---

Citation

@dataset{hdx_africa_cod_ps_global,
  title     = {OCHA Global Subnational Population Statistics},
  author    = {HDX},
  year      = {2026},
  url       = {https://data.humdata.org/dataset/cod-ps-global},
  note      = {Repackaged for machine learning by Electric Sheep Africa (https://huggingface.co/electricsheepafrica)}
}

---

Electric Sheep Africa — Africa's ML dataset infrastructure. Lagos, Nigeria.