electricsheepasia/asia-population-emdat-country-profiles-pakistan

EM-DAT - Country Profiles, Pakistan

Publisher: Centre for Research on the Epidemiology of Disasters · Source: HDX · License: hdx-other · Updated: 2026-05-02

---

Abstract

Aggregated figures for natural hazard related events in EM-DAT: Pakistan

Documentation on the Country Profiles available here

How to cite the EM-DAT Project here

Main dataset on HDX: EM-DAT - Country Profiles

More on the EM-DAT database : website / data portal

Each line corresponds to a given combination of year, country, disaster subtype and reports figures for :

• number of disasters
• total number of people affected
• total number of deaths
• economic losses (original value and adjusted)

---

Each row in this dataset represents country-level aggregates. Data was last updated on HDX on 2026-05-02. Geographic scope: PAK.

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

---

Dataset Characteristics

Domain	Demographics and population
Unit of observation	Country-level aggregates
Rows (total)	108
Columns	15 (7 numeric, 8 categorical, 0 datetime)
Train split	86 rows
Test split	21 rows
Geographic scope	PAK
Publisher	Centre for Research on the Epidemiology of Disasters
HDX last updated	2026-05-02

---

Variables

Geographic — year (range 2000.0–2026.0), country (Pakistan, #country +name), iso (PAK, #country +code), disaster_type (Flood, Storm, Earthquake), disaster_subtype (Flash flood, Riverine flood, Ground movement).

Demographic — total_damage_usd_original (range 30000.0–15000000000.0), total_damage_usd_adjusted (range 52318.0–16078099716.0).

Outcome / Measurement — total_events (range 1.0–6.0), total_affected (range 2.0–33012865.0), total_deaths (range 1.0–73338.0).

Identifier / Metadata — esa_source (HDX), esa_processed (2026-05-06).

Other — disaster_group (Natural, #cause +group), disaster_subroup (Hydrological, Meteorological, Geophysical), cpi (range 54.8952–100.0).

---

Quick Start

from datasets import load_dataset

ds    = load_dataset("electricsheepafrica/asia-population-emdat-country-profiles-pakistan")
train = ds["train"].to_pandas()
test  = ds["test"].to_pandas()

print(train.shape)
train.head()

---

Schema

Column	Type	Null %	Range / Sample Values
year	float64	0.9%	2000.0 – 2026.0 (mean 2012.7664)
country	object	0.0%	Pakistan, #country +name
iso	object	0.0%	PAK, #country +code
disaster_group	object	0.0%	Natural, #cause +group
disaster_subroup	object	0.0%	Hydrological, Meteorological, Geophysical
disaster_type	object	0.0%	Flood, Storm, Earthquake
disaster_subtype	object	0.0%	Flash flood, Riverine flood, Ground movement
total_events	float64	0.9%	1.0 – 6.0 (mean 1.514)
total_affected	float64	25.9%	2.0 – 33012865.0 (mean 1286162.8875)
total_deaths	float64	3.7%	1.0 – 73338.0 (mean 864.7885)
total_damage_usd_original	float64	75.0%	30000.0 – 15000000000.0 (mean 1681394370.3704)
total_damage_usd_adjusted	float64	76.9%	52318.0 – 16078099716.0 (mean 2222366629.24)
cpi	float64	5.6%	54.8952 – 100.0 (mean 73.5271)
esa_source	object	0.0%	HDX
esa_processed	object	0.0%	2026-05-06

---

Numeric Summary

Column	Min	Max	Mean	Median
year	2000.0	2026.0	2012.7664	2013.0
total_events	1.0	6.0	1.514	1.0
total_affected	2.0	33012865.0	1286162.8875	3852.5
total_deaths	1.0	73338.0	864.7885	50.5
total_damage_usd_original	30000.0	15000000000.0	1681394370.3704	103000000.0
total_damage_usd_adjusted	52318.0	16078099716.0	2222366629.24	140771167.0
cpi	54.8952	100.0	73.5271	73.1916

---

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. 5 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 Centre for Research on the Epidemiology of Disasters 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: total_affected, total_damage_usd_original, total_damage_usd_adjusted.
• Refer to the original HDX dataset page for the publisher's own methodology notes and caveats.

---

Citation

@dataset{hdx_asia_population_emdat_country_profiles_pakistan,
  title     = {EM-DAT - Country Profiles, Pakistan},
  author    = {Centre for Research on the Epidemiology of Disasters},
  year      = {2026},
  url       = {https://data.humdata.org/dataset/emdat-country-profiles-pak},
  note      = {Repackaged for machine learning by Electric Sheep Africa (https://huggingface.co/electricsheepafrica)}
}

---

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