electricsheepafrica/africa-unhcr-population-data-for-som

Somalia - Data on forcibly displaced populations and stateless persons

Publisher: UNHCR - The UN Refugee Agency · Source: HDX · License: cc-by-igo · Updated: 2026-02-25

---

Abstract

Data collated by UNHCR, containing information about forcibly displaced populations and stateless persons, spanning across more than 70 years of statistical activities. The data includes the countries / territories of asylum and origin. Specific resources are available for end-year population totals, demographics, asylum applications, decisions, and solutions availed by refugees and IDPs (resettlement, naturalisation or returns).

Each row in this dataset represents first-level administrative unit observations. Data was last updated on HDX on 2026-02-25. Geographic scope: SOM.

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

---

Dataset Characteristics

Domain	Demographics and population
Unit of observation	First-level administrative unit observations
Rows (total)	3,035
Columns	14 (8 numeric, 6 categorical, 0 datetime)
Train split	2,428 rows
Test split	607 rows
Geographic scope	SOM
Publisher	UNHCR - The UN Refugee Agency
HDX last updated	2026-02-25

---

Variables

Geographic — year (range 1975.0–2025.0), country_of_origin_code (SOM), country_of_asylum_code (KEN, ITA, ETH), country_of_origin_name (Somalia), country_of_asylum_name (Kenya, Italy, Ethiopia) and 4 others.

Identifier / Metadata — refugees (range 0.0–517666.0), esa_source (HDX), esa_processed (2026-04-04).

Other — other_people_in_need_of_international_protection (range 0.0–0.0), others_of_concern_to_unhcr (range 0.0–85820.0).

---

Quick Start

from datasets import load_dataset

ds    = load_dataset("electricsheepafrica/africa-unhcr-population-data-for-som")
train = ds["train"].to_pandas()
test  = ds["test"].to_pandas()

print(train.shape)
train.head()

---

Schema

Column	Type	Null %	Range / Sample Values
year	int64	0.0%	1975.0 – 2025.0 (mean 2009.9908)
country_of_origin_code	object	0.0%	SOM
country_of_asylum_code	object	0.0%	KEN, ITA, ETH
country_of_origin_name	object	0.0%	Somalia
country_of_asylum_name	object	0.0%	Kenya, Italy, Ethiopia
refugees	int64	0.0%	0.0 – 517666.0 (mean 8851.1367)
asylum_seekers	int64	0.0%	0.0 – 150306.0 (mean 486.1516)
other_people_in_need_of_international_protection	int64	0.0%	0.0 – 0.0 (mean 0.0)
internally_displaced_persons	int64	0.0%	0.0 – 3861643.0 (mean 13673.1776)
stateless_persons	int64	0.0%	0.0 – 0.0 (mean 0.0)
others_of_concern_to_unhcr	int64	0.0%	0.0 – 85820.0 (mean 100.8326)
host_community	int64	0.0%	0.0 – 0.0 (mean 0.0)
esa_source	object	0.0%	HDX
esa_processed	object	0.0%	2026-04-04

---

Numeric Summary

Column	Min	Max	Mean	Median
year	1975.0	2025.0	2009.9908	2011.0
refugees	0.0	517666.0	8851.1367	67.0
asylum_seekers	0.0	150306.0	486.1516	5.0
other_people_in_need_of_international_protection	0.0	0.0	0.0	0.0
internally_displaced_persons	0.0	3861643.0	13673.1776	0.0
stateless_persons	0.0	0.0	0.0	0.0
others_of_concern_to_unhcr	0.0	85820.0	100.8326	0.0
host_community	0.0	0.0	0.0	0.0

---

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. 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 UNHCR - The UN Refugee Agency and has not been independently validated by ESA.
• Automated cleaning cannot correct for misreported values, definitional inconsistencies, or sampling bias in the original collection.
• Refer to the original HDX dataset page for the publisher's own methodology notes and caveats.

---

Citation

@dataset{hdx_africa_unhcr_population_data_for_som,
  title     = {Somalia - Data on forcibly displaced populations and stateless persons},
  author    = {UNHCR - The UN Refugee Agency},
  year      = {2026},
  url       = {https://data.humdata.org/dataset/unhcr-population-data-for-som},
  note      = {Repackaged for machine learning by Electric Sheep Africa (https://huggingface.co/electricsheepafrica)}
}

---

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