electricsheepafrica/africa-global-mpi

---
annotations_creators:
- no-annotation
language_creators:
- found
language:
- en
license: other
multilinguality:
- monolingual
size_categories:
- 1K<n<10K
source_datasets:
- original
task_categories:
- tabular-classification
- tabular-regression
- other
task_ids: []
tags:
- africa
- humanitarian
- hdx
- electric-sheep-africa
- development
- education
- health
- indicators
- mortality
- nutrition
- poverty
- socioeconomics
- afg
- alb
- dza
- ago
- arg
pretty_name: "Global Multidimensional Poverty Index"
dataset_info:
splits:
- name: train
num_examples: 1174
- name: test
num_examples: 293
---

# Global Multidimensional Poverty Index

**Publisher:** Oxford Poverty & Human Development Initiative · **Source:** [HDX](https://data.humdata.org/dataset/global-mpi) · **License:** `other-pd-nr` · **Updated:** 2026-03-05

---

## Abstract

The global Multidimensional Poverty Index provides the only comprehensive measure available for non-income poverty, which has become a critical underpinning of the SDGs. The global Multidimensional Poverty Index (MPI) measures multidimensional poverty in over 100 developing countries, using internationally comparable datasets and is updated annually. The measure captures the acute deprivations that each person faces at the same time using information from 10 indicators, which are grouped into three equally weighted dimensions: health, education, and living standards. Critically, the MPI comprises variables that are already reported under the Demographic Health Surveys (DHS), the Multi-Indicator Cluster Surveys (MICS) and in some cases, national surveys.

The subnational multidimensional poverty data from the [data tables](https://ophi.org.uk/global-mpi-archive) are published by the Oxford Poverty and Human Development Initiative (OPHI), University of Oxford. For the details of the global MPI methodology, please see the latest Methodological Notes [found here](https://ophi.org.uk/publications-table?title=&field_authors_value=&field_publication_type_target_id=11&publication_year_filter=All&field_keywords_value=&field_country_target_id=All&field_region_target_id=All).

Each row in this dataset represents country-level aggregates. Data was last updated on HDX on 2026-03-05. Geographic scope: **AFG, ALB, DZA, AGO, ARG, ARM, BGD, BRB, and 104 others**.

*Curated into ML-ready Parquet format by [Electric Sheep Africa](https://huggingface.co/electricsheepafrica).*

---

## Dataset Characteristics

| | |
|---|---|
| **Domain** | Public health |
| **Unit of observation** | Country-level aggregates |
| **Rows (total)** | 1,468 |
| **Columns** | 13 (5 numeric, 6 categorical, 0 datetime) |
| **Train split** | 1,174 rows |
| **Test split** | 293 rows |
| **Geographic scope** | AFG, ALB, DZA, AGO, ARG, ARM, BGD, BRB, and 104 others |
| **Publisher** | Oxford Poverty & Human Development Initiative |
| **HDX last updated** | 2026-03-05 |

---

## Variables

**Geographic** — `country_iso3` (KEN, NGA, IND), `admin_1_pcode` (HT04, TN3, GH08), `admin_1_name` (Central, Western, Eastern), `intensity_of_deprivation` (range 33.3333–69.6577), `vulnerable_to_poverty` (range 0.0–48.3534) and 2 others.

**Temporal** — `start_date`, `end_date`.

**Outcome / Measurement** — `headcount_ratio` (range 0.0–99.2537).

**Identifier / Metadata** — `esa_source` (HDX), `esa_processed` (2026-04-04).

**Other** — `mpi` (range 0.0–0.6759).

---

## Quick Start

```python
from datasets import load_dataset

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

print(train.shape)
train.head()
```

---

## Schema

| Column | Type | Null % | Range / Sample Values |
|---|---|---|---|
| `country_iso3` | object | 0.0% | KEN, NGA, IND |
| `admin_1_pcode` | object | 40.7% | HT04, TN3, GH08 |
| `admin_1_name` | object | 7.4% | Central, Western, Eastern |
| `mpi` | float64 | 0.0% | 0.0 – 0.6759 (mean 0.1482) |
| `headcount_ratio` | float64 | 0.0% | 0.0 – 99.2537 (mean 29.2917) |
| `intensity_of_deprivation` | float64 | 1.2% | 33.3333 – 69.6577 (mean 44.2109) |
| `vulnerable_to_poverty` | float64 | 0.0% | 0.0 – 48.3534 (mean 15.0861) |
| `in_severe_poverty` | float64 | 0.0% | 0.0 – 91.6007 (mean 14.5411) |
| `survey` | object | 0.0% | DHS, MICS, PAPFAM |
| `start_date` | datetime64[ns, UTC] | 0.0% | |
| `end_date` | datetime64[ns, UTC] | 0.0% | |
| `esa_source` | object | 0.0% | HDX |
| `esa_processed` | object | 0.0% | 2026-04-04 |

---

## Numeric Summary

| Column | Min | Max | Mean | Median |
|---|---|---|---|---|
| `mpi` | 0.0 | 0.6759 | 0.1482 | 0.0743 |
| `headcount_ratio` | 0.0 | 99.2537 | 29.2917 | 17.418 |
| `intensity_of_deprivation` | 33.3333 | 69.6577 | 44.2109 | 42.5692 |
| `vulnerable_to_poverty` | 0.0 | 48.3534 | 15.0861 | 14.8096 |
| `in_severe_poverty` | 0.0 | 91.6007 | 14.5411 | 3.8831 |

---

## 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`. 2 column(s) were cast from string to numeric or datetime based on parse-success rate (>85% threshold). 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 Oxford Poverty & Human Development Initiative and has not been independently validated by ESA.
- Automated cleaning cannot correct for misreported values, definitional inconsistencies, or sampling bias in the original collection.
- The following columns have >20% missing values and should be treated with caution in modelling: `admin_1_pcode`.
- This dataset spans 112 countries; geographic and methodological inconsistencies across national boundaries may affect cross-country comparability.
- Refer to the [original HDX dataset page](https://data.humdata.org/dataset/global-mpi) for the publisher's own methodology notes and caveats.

---

## Citation

```bibtex
@dataset{hdx_africa_global_mpi,
title = {Global Multidimensional Poverty Index},
author = {Oxford Poverty & Human Development Initiative},
year = {2026},
url = {https://data.humdata.org/dataset/global-mpi},
note = {Repackaged for machine learning by Electric Sheep Africa (https://huggingface.co/electricsheepafrica)}
}
```

---

*[Electric Sheep Africa](https://huggingface.co/electricsheepafrica) — Africa's ML dataset infrastructure. Lagos, Nigeria.*