electricsheepafrica/africa-demographics-eswatini

---
annotations_creators:
- no-annotation
language_creators:
- found
language:
- en
license: other
multilinguality:
- monolingual
size_categories:
- n<1K
source_datasets:
- original
task_categories:
- tabular-classification
- other
task_ids: []
tags:
- africa
- humanitarian
- hdx
- electric-sheep-africa
- demographics
- health
- swz
pretty_name: "Eswatini - Subnational Demographic and Health Data"
dataset_info:
splits:
- name: train
num_examples: 89
- name: test
num_examples: 22
---

# Eswatini - Subnational Demographic and Health Data

**Publisher:** The DHS Program · **Source:** [HDX](https://data.humdata.org/dataset/dhs-subnational-data-for-eswatini) · **License:** `hdx-other` · **Updated:** 2026-04-20

---

## Abstract

Contains data from the [DHS data portal](https://api.dhsprogram.com/). There is also a dataset containing [Eswatini - National Demographic and Health Data](https://data.humdata.org/dataset/dhs-data-for-eswatini) on HDX.

The DHS Program Application Programming Interface (API) provides software developers access to aggregated indicator data from The Demographic and Health Surveys (DHS) Program. The API can be used to create various applications to help analyze, visualize, explore and disseminate data on population, health, HIV, and nutrition from more than 90 countries.

Each row in this dataset represents first-level administrative unit observations. Temporal coverage is indicated by the `surveyyearlabel` column(s). Geographic scope: **SWZ**.

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

---

## Dataset Characteristics

| | |
|---|---|
| **Domain** | Public health |
| **Unit of observation** | First-level administrative unit observations |
| **Rows (total)** | 112 |
| **Columns** | 30 (13 numeric, 16 categorical, 1 datetime) |
| **Train split** | 89 rows |
| **Test split** | 22 rows |
| **Geographic scope** | SWZ |
| **Publisher** | The DHS Program |
| **HDX last updated** | 2026-04-20 |

---

## Variables

**Geographic** — `iso3` (SWZ), `location` (Hhohho, Manzini, Shiselweni), `dhs_countrycode` (SZ), `countryname` (Eswatini), `surveyyear` (range 2006.0–2006.0) and 8 others.

**Outcome / Measurement** — `value` (range 0.0–114.0), `istotal` (range 0.0–0.0).

**Identifier / Metadata** — `dataid` (range 239470.0–7981495.0), `indicatorid` (RH_DELP_C_DHF, CH_DIAT_C_ORT, FE_FRTR_W_TFR), `characteristicid` (range 502001.0–502004.0), `characteristiclabel` (Hhohho, Manzini, Shiselweni), `ispreferred` (range 0.0–1.0) and 3 others.

**Other** — `indicator` (Place of delivery: Health facility, Treatment of diarrhea: Either ORS or RHF, Total fertility rate 15-49), `precision` (range 0.0–1.0), `indicatororder` (range 11763080.0–260321010.0), `characteristicorder` (range 1502001.0–1502004.0), `denominatorweighted` (range 66.0–2682.0) and 2 others.

---

## Quick Start

```python
from datasets import load_dataset

ds = load_dataset("electricsheepafrica/africa-demographics-eswatini")
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% | SWZ |
| `location` | object | 0.0% | Hhohho, Manzini, Shiselweni |
| `dataid` | int64 | 0.0% | 239470.0 – 7981495.0 (mean 4319947.5179) |
| `indicator` | object | 0.0% | Place of delivery: Health facility, Treatment of diarrhea: Either ORS or RHF, Total fertility rate 15-49 |
| `value` | float64 | 0.0% | 0.0 – 114.0 (mean 43.5813) |
| `precision` | int64 | 0.0% | 0.0 – 1.0 (mean 0.9286) |
| `dhs_countrycode` | object | 0.0% | SZ |
| `countryname` | object | 0.0% | Eswatini |
| `surveyyear` | int64 | 0.0% | 2006.0 – 2006.0 (mean 2006.0) |
| `surveyid` | object | 0.0% | SZ2006DHS |
| `indicatorid` | object | 0.0% | RH_DELP_C_DHF, CH_DIAT_C_ORT, FE_FRTR_W_TFR |
| `indicatororder` | int64 | 0.0% | 11763080.0 – 260321010.0 (mean 101396331.0714) |
| `indicatortype` | object | 0.0% | I |
| `characteristicid` | int64 | 0.0% | 502001.0 – 502004.0 (mean 502002.5) |
| `characteristicorder` | int64 | 0.0% | 1502001.0 – 1502004.0 (mean 1502002.5) |
| `characteristiccategory` | object | 0.0% | Region |
| `characteristiclabel` | object | 0.0% | Hhohho, Manzini, Shiselweni |
| `byvariableid` | int64 | 0.0% | 0.0 – 14003.0 (mean 3500.3571) |
| `byvariablelabel` | object | 75.0% | |
| `istotal` | int64 | 0.0% | 0.0 – 0.0 (mean 0.0) |
| `ispreferred` | int64 | 0.0% | 0.0 – 1.0 (mean 0.8929) |
| `sdrid` | object | 0.0% | |
| `regionid` | object | 0.0% | |
| `surveyyearlabel` | datetime64[ns] | 0.0% | |
| `surveytype` | object | 0.0% | |
| `denominatorweighted` | float64 | 21.4% | 66.0 – 2682.0 (mean 762.4886) |
| `denominatorunweighted` | float64 | 21.4% | 58.0 – 2362.0 (mean 755.0341) |
| `levelrank` | int64 | 0.0% | 1.0 – 1.0 (mean 1.0) |
| `esa_source` | object | 0.0% | |
| `esa_processed` | object | 0.0% | |

---

## Numeric Summary

| Column | Min | Max | Mean | Median |
|---|---|---|---|---|
| `dataid` | 239470.0 | 7981495.0 | 4319947.5179 | 4366443.5 |
| `value` | 0.0 | 114.0 | 43.5813 | 32.2 |
| `precision` | 0.0 | 1.0 | 0.9286 | 1.0 |
| `surveyyear` | 2006.0 | 2006.0 | 2006.0 | 2006.0 |
| `indicatororder` | 11763080.0 | 260321010.0 | 101396331.0714 | 94096040.0 |
| `characteristicid` | 502001.0 | 502004.0 | 502002.5 | 502002.5 |
| `characteristicorder` | 1502001.0 | 1502004.0 | 1502002.5 | 1502002.5 |
| `byvariableid` | 0.0 | 14003.0 | 3500.3571 | 0.0 |
| `istotal` | 0.0 | 0.0 | 0.0 | 0.0 |
| `ispreferred` | 0.0 | 1.0 | 0.8929 | 1.0 |
| `denominatorweighted` | 66.0 | 2682.0 | 762.4886 | 745.0 |
| `denominatorunweighted` | 58.0 | 2362.0 | 755.0341 | 694.0 |
| `levelrank` | 1.0 | 1.0 | 1.0 | 1.0 |

---

## 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`. 2 column(s) with >80% missing values were removed: `cilow`, `cihigh`. 1 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 The DHS Program 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: `byvariablelabel`, `denominatorweighted`, `denominatorunweighted`.
- Refer to the [original HDX dataset page](https://data.humdata.org/dataset/dhs-subnational-data-for-eswatini) for the publisher's own methodology notes and caveats.

---

## Citation

```bibtex
@dataset{hdx_africa_demographics_eswatini,
title = {Eswatini - Subnational Demographic and Health Data},
author = {The DHS Program},
year = {2026},
url = {https://data.humdata.org/dataset/dhs-subnational-data-for-eswatini},
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.*