electricsheepafrica/africa-ucdp-data-for-eswatini

---
annotations_creators:
- no-annotation
language_creators:
- found
language:
- en
license: cc-by-4.0
multilinguality:
- monolingual
size_categories:
- n<1K
source_datasets:
- original
task_categories:
- tabular-classification
- other
task_ids: []
tags:
- africa
- humanitarian
- hdx
- electric-sheep-africa
- conflict-violence
- hxl
- swz
pretty_name: "Eswatini - Data on Conflict Events"
dataset_info:
splits:
- name: train
num_examples: 17
- name: test
num_examples: 4
---

# Eswatini - Data on Conflict Events

**Publisher:** HDX · **Source:** [HDX](https://data.humdata.org/dataset/ucdp-data-for-eswatini) · **License:** `cc-by-igo` · **Updated:** 2026-04-03

---

## Abstract

This dataset is UCDP's most disaggregated dataset, covering individual events of organized violence (phenomena of lethal violence occurring at a given time and place). These events are sufficiently fine-grained to be geo-coded down to the level of individual villages, with temporal durations disaggregated to single, individual days.
Sundberg, Ralph, and Erik Melander, 2013, “Introducing the UCDP Georeferenced Event Dataset”, Journal of Peace Research, vol.50, no.4, 523-532
Högbladh Stina, 2019, “UCDP GED Codebook version 19.1”, Department of Peace and Conflict Research, Uppsala University

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

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

---

## Dataset Characteristics

| | |
|---|---|
| **Domain** | Conflict and security |
| **Unit of observation** | First-level administrative unit observations |
| **Rows (total)** | 22 |
| **Columns** | 51 (27 numeric, 21 categorical, 2 datetime) |
| **Train split** | 17 rows |
| **Test split** | 4 rows |
| **Geographic scope** | SWZ |
| **Publisher** | HDX |
| **HDX last updated** | 2026-04-03 |

---

## Variables

**Geographic** — `year` (range 1989.0–2023.0), `active_year`, `type_of_violence` (range 3.0–3.0), `dyad_dset_id` (range 102.0–1212.0), `dyad_new_id` (range 944.0–2485.0) and 9 others.

**Temporal** — `source_date` (2020-08-04, 2020-04-09, 2021-08-12;2021-07-02;2021-09-05), `date_prec` (range 1.0–4.0), `date_start`, `date_end`.

**Outcome / Measurement** — `number_of_sources` (range -1.0–5.0), `deaths_a` (range 0.0–0.0), `deaths_b`, `deaths_civilians`, `deaths_unknown`.

**Identifier / Metadata** — `id` (range 14863.0–502824.0), `relid` (SAF-1989-3-560-31, SAF-1989-3-560-5, SWA-2023-3-2485-0), `code_status` (Clear), `conflict_dset_id` (range 102.0–1212.0), `conflict_new_id` (range 477.0–2003.0) and 14 others.

**Other** — `where_prec` (range 1.0–6.0), `where_description`, `adm_1`, `adm_2`, `geom_wkt` and 4 others.

---

## Quick Start

```python
from datasets import load_dataset

ds = load_dataset("electricsheepafrica/africa-ucdp-data-for-eswatini")
train = ds["train"].to_pandas()
test = ds["test"].to_pandas()

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

---

## Schema

| Column | Type | Null % | Range / Sample Values |
|---|---|---|---|
| `id` | int64 | 0.0% | 14863.0 – 502824.0 (mean 371260.2273) |
| `relid` | object | 0.0% | SAF-1989-3-560-31, SAF-1989-3-560-5, SWA-2023-3-2485-0 |
| `year` | int64 | 0.0% | 1989.0 – 2023.0 (mean 2018.0) |
| `active_year` | bool | 0.0% | |
| `code_status` | object | 0.0% | Clear |
| `type_of_violence` | int64 | 0.0% | 3.0 – 3.0 (mean 3.0) |
| `conflict_dset_id` | int64 | 0.0% | 102.0 – 1212.0 (mean 1111.0909) |
| `conflict_new_id` | int64 | 0.0% | 477.0 – 2003.0 (mean 1864.2727) |
| `conflict_name` | object | 0.0% | Government of eSwatini (Swaziland) - Civilians, Government of South Africa - Civilians |
| `dyad_dset_id` | int64 | 0.0% | 102.0 – 1212.0 (mean 1111.0909) |
| `dyad_new_id` | int64 | 0.0% | 944.0 – 2485.0 (mean 2344.9091) |
| `dyad_name` | object | 0.0% | Government of eSwatini (Swaziland) - Civilians, Government of South Africa - Civilians |
| `side_a_dset_id` | int64 | 0.0% | 102.0 – 1212.0 (mean 1111.0909) |
| `side_a_new_id` | int64 | 0.0% | 102.0 – 1212.0 (mean 1111.0909) |
| `side_a` | object | 0.0% | Government of eSwatini (Swaziland), Government of South Africa |
| `side_b_dset_id` | int64 | 0.0% | 9999.0 – 9999.0 (mean 9999.0) |
| `side_b_new_id` | int64 | 0.0% | 1.0 – 1.0 (mean 1.0) |
| `side_b` | object | 0.0% | Civilians |
| `number_of_sources` | int64 | 0.0% | -1.0 – 5.0 (mean 2.0455) |
| `source_article` | object | 0.0% | The report of the Truth and Reconciliation Commission (TRC report) https://www.justice.gov.za/trc/report/, "All Africa,2020-08-04,Swaziland Police Kill Another Unarmed Man Reviving Fears of Shoot-to-Kill Policy
CR Swazi Media Commentary", "New Frame ,2021-08-12,Long Read | eSwatini: the people versus the king";"Amnesty International,2021-07-02,Eswatini: Dozens killed, tortured, abducted as pro-democracy protests intensify";"Pudemo's Twitter account,2021-09-05,THEY HAVE NAMES." |
| `source_office` | object | 9.1% | All Africa, Times of Swaziland, Agence France Presse |
| `source_date` | object | 9.1% | 2020-08-04, 2020-04-09, 2021-08-12;2021-07-02;2021-09-05 |
| `source_headline` | object | 9.1% | Swaziland Police Kill Another Unarmed Man Reviving Fears of Shoot-to-Kill Policy
CR Swazi Media Commentary, Swaziland Govt or Agents 'Committed Unlawful Killings', New Human Rights Report States
CR Swazi Media Commentary, Long Read | eSwatini: the people versus the king;Eswatini: Dozens killed, tortured, abducted as pro-democracy protests intensify;THEY HAVE NAMES. |
| `source_original` | object | 18.2% | |
| `where_prec` | int64 | 0.0% | 1.0 – 6.0 (mean 2.4545) |
| `where_coordinates` | object | 0.0% | |
| `where_description` | object | 18.2% | |
| `adm_1` | object | 22.7% | |
| `adm_2` | object | 59.1% | |
| `latitude` | float64 | 0.0% | -27.253 – -26.2941 (mean -26.518) |
| `longitude` | float64 | 0.0% | 30.9229 – 31.7686 (mean 31.3021) |
| `geom_wkt` | object | 0.0% | |
| `priogrid_gid` | int64 | 0.0% | 90423.0 – 91864.0 (mean 91601.4545) |
| `country` | object | 0.0% | |
| `iso3` | object | 0.0% | |
| `country_id` | int64 | 0.0% | 572.0 – 572.0 (mean 572.0) |
| `region` | object | 0.0% | |
| `event_clarity` | int64 | 0.0% | 1.0 – 2.0 (mean 1.0909) |
| `date_prec` | int64 | 0.0% | 1.0 – 4.0 (mean 1.3182) |
| `date_start` | datetime64[ns] | 0.0% | |
| `date_end` | datetime64[ns] | 0.0% | |
| `deaths_a` | int64 | 0.0% | 0.0 – 0.0 (mean 0.0) |
| `deaths_b` | int64 | 0.0% | |
| `deaths_civilians` | int64 | 0.0% | |
| `deaths_unknown` | int64 | 0.0% | |
| `best` | int64 | 0.0% | |
| `high` | int64 | 0.0% | |
| `low` | int64 | 0.0% | |
| `gwnoa` | int64 | 0.0% | |
| `esa_source` | object | 0.0% | |
| `esa_processed` | object | 0.0% | |

---

## Numeric Summary

| Column | Min | Max | Mean | Median |
|---|---|---|---|---|
| `id` | 14863.0 | 502824.0 | 371260.2273 | 410600.0 |
| `year` | 1989.0 | 2023.0 | 2018.0 | 2021.0 |
| `type_of_violence` | 3.0 | 3.0 | 3.0 | 3.0 |
| `conflict_dset_id` | 102.0 | 1212.0 | 1111.0909 | 1212.0 |
| `conflict_new_id` | 477.0 | 2003.0 | 1864.2727 | 2003.0 |
| `dyad_dset_id` | 102.0 | 1212.0 | 1111.0909 | 1212.0 |
| `dyad_new_id` | 944.0 | 2485.0 | 2344.9091 | 2485.0 |
| `side_a_dset_id` | 102.0 | 1212.0 | 1111.0909 | 1212.0 |
| `side_a_new_id` | 102.0 | 1212.0 | 1111.0909 | 1212.0 |
| `side_b_dset_id` | 9999.0 | 9999.0 | 9999.0 | 9999.0 |
| `side_b_new_id` | 1.0 | 1.0 | 1.0 | 1.0 |
| `number_of_sources` | -1.0 | 5.0 | 2.0455 | 1.0 |
| `where_prec` | 1.0 | 6.0 | 2.4545 | 2.0 |
| `latitude` | -27.253 | -26.2941 | -26.518 | -26.4833 |
| `longitude` | 30.9229 | 31.7686 | 31.3021 | 31.2662 |

---

## 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`. 1 column(s) with >80% missing values were removed: `gwnob`. 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 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.
- The following columns have >20% missing values and should be treated with caution in modelling: `adm_1`, `adm_2`.
- Refer to the [original HDX dataset page](https://data.humdata.org/dataset/ucdp-data-for-eswatini) for the publisher's own methodology notes and caveats.

---

## Citation

```bibtex
@dataset{hdx_africa_ucdp_data_for_eswatini,
title = {Eswatini - Data on Conflict Events},
author = {HDX},
year = {2026},
url = {https://data.humdata.org/dataset/ucdp-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.*