electricsheepafrica/africa-ucdp-data-for-nigeria

Nigeria - Data on Conflict Events

Publisher: HDX · Source: HDX · 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: NGA.

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

---

Dataset Characteristics

Domain	Conflict and security
Unit of observation	First-level administrative unit observations
Rows (total)	7,418
Columns	51 (27 numeric, 21 categorical, 2 datetime)
Train split	5,934 rows
Test split	1,483 rows
Geographic scope	NGA
Publisher	HDX
HDX last updated	2026-04-03

---

Variables

Geographic — year (range 1990.0–2024.0), active_year, type_of_violence (range 1.0–3.0), dyad_dset_id (range 83.0–18023.0), dyad_new_id (range 640.0–18092.0) and 9 others.

Temporal — source_date (2013-12-01, 2015-04-30, 2023-01-26), date_prec (range 1.0–5.0), date_start, date_end.

Outcome / Measurement — number_of_sources (range -1.0–25.0), deaths_a (range 0.0–200.0), deaths_b, deaths_civilians, deaths_unknown.

Identifier / Metadata — id (range 13223.0–568733.0), relid (NIG-2009-1-793-1, NIG-2015-1-14669-20, NIG-2015-1-14669-18), code_status (Clear), conflict_dset_id (range 83.0–18023.0), conflict_new_id (range 297.0–16330.0) and 14 others.

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

---

Quick Start

from datasets import load_dataset

ds    = load_dataset("electricsheepafrica/africa-ucdp-data-for-nigeria")
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%	13223.0 – 568733.0 (mean 265727.2277)
relid	object	0.0%	NIG-2009-1-793-1, NIG-2015-1-14669-20, NIG-2015-1-14669-18
year	int64	0.0%	1990.0 – 2024.0 (mean 2016.7803)
active_year	bool	0.0%
code_status	object	0.0%	Clear
type_of_violence	int64	0.0%	1.0 – 3.0 (mean 1.7913)
conflict_dset_id	int64	0.0%	83.0 – 18023.0 (mean 6775.0023)
conflict_new_id	int64	0.0%	297.0 – 16330.0 (mean 6718.2467)
conflict_name	object	0.0%	Nigeria: Islamic State, Nigeria: Government, JAS - Civilians
dyad_dset_id	int64	0.0%	83.0 – 18023.0 (mean 7507.2892)
dyad_new_id	int64	0.0%	640.0 – 18092.0 (mean 7816.6917)
dyad_name	object	0.0%	Government of Nigeria - IS, Government of Nigeria - JAS, JAS - Civilians
side_a_dset_id	int64	0.0%	83.0 – 9195.0 (mean 741.051)
side_a_new_id	int64	0.0%	83.0 – 9195.0 (mean 741.051)
side_a	object	0.0%	Government of Nigeria, JAS, Fulani
side_b_dset_id	int64	0.0%	234.0 – 9999.0 (mean 3392.244)
side_b_new_id	int64	0.0%	1.0 – 9124.0 (mean 1078.0669)
side_b	object	0.0%	IS, Civilians, JAS
number_of_sources	int64	0.0%	-1.0 – 25.0 (mean 1.3318)
source_article	object	0.0%	NigeriaWatch database, "Zagazola,2023-01-26,Troops Kill 157 Terrorists as NAF Jets’ Destroy ISWAP Enclaves in Borno", "All Africa,2017-03-14,Boko Haram - Troops Rescue 455 Captives, Kill Scores of Terrorists"
source_office	object	16.0%	All Africa, Zagazola, Agence France Presse
source_date	object	16.0%	2013-12-01, 2015-04-30, 2023-01-26
source_headline	object	16.0%	Situation North East (Adamawa, Borno and Yobe States), Boko Haram Attack, Borno, Army vs Boko Haram , Borno
source_original	object	25.0%
where_prec	int64	0.0%	1.0 – 6.0 (mean 1.9229)
where_coordinates	object	0.0%
where_description	object	0.4%
adm_1	object	0.6%
adm_2	object	13.2%
latitude	float64	0.0%	4.2864 – 13.6911 (mean 10.1301)
longitude	float64	0.0%	2.7104 – 14.5 (mean 10.8979)
geom_wkt	object	0.0%
priogrid_gid	int64	0.0%	135733.0 – 149428.0 (mean 144206.2416)
country	object	0.0%
iso3	object	0.0%
country_id	int64	0.0%	475.0 – 475.0 (mean 475.0)
region	object	0.0%
event_clarity	int64	0.0%	1.0 – 2.0 (mean 1.0861)
date_prec	int64	0.0%	1.0 – 5.0 (mean 1.2773)
date_start	datetime64[ns]	0.0%
date_end	datetime64[ns]	0.0%
deaths_a	int64	0.0%	0.0 – 200.0 (mean 1.1971)
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	float64	51.1%
esa_source	object	0.0%
esa_processed	object	0.0%

---

Numeric Summary

Column	Min	Max	Mean	Median
id	13223.0	568733.0	265727.2277	254373.5
year	1990.0	2024.0	2016.7803	2018.0
type_of_violence	1.0	3.0	1.7913	2.0
conflict_dset_id	83.0	18023.0	6775.0023	5558.0
conflict_new_id	297.0	16330.0	6718.2467	4948.0
dyad_dset_id	83.0	18023.0	7507.2892	5583.0
dyad_new_id	640.0	18092.0	7816.6917	5583.0
side_a_dset_id	83.0	9195.0	741.051	234.0
side_a_new_id	83.0	9195.0	741.051	234.0
side_b_dset_id	234.0	9999.0	3392.244	1051.0
side_b_new_id	1.0	9124.0	1078.0669	607.0
number_of_sources	-1.0	25.0	1.3318	1.0
where_prec	1.0	6.0	1.9229	1.0
latitude	4.2864	13.6911	10.1301	11.1036
longitude	2.7104	14.5	10.8979	12.4916

---

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: source_original, gwnoa.
• Refer to the original HDX dataset page for the publisher's own methodology notes and caveats.

---

Citation

@dataset{hdx_africa_ucdp_data_for_nigeria,
  title     = {Nigeria - Data on Conflict Events},
  author    = {HDX},
  year      = {2026},
  url       = {https://data.humdata.org/dataset/ucdp-data-for-nigeria},
  note      = {Repackaged for machine learning by Electric Sheep Africa (https://huggingface.co/electricsheepafrica)}
}

---

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