electricsheepafrica/africa-icpac-geonode-2015-tropical-cyclone-megh-path

2015 Tropical Cyclone Megh Path

Publisher: IGAD Climate Prediction and Applications Center (ICPAC) · Source: HDX · License: cc-by · Updated: 2026-04-05

---

Abstract

This layer shows the movement path for 2015 Tropical Cyclone Megh. Following Tropical Cyclone Chapala, new tropical cyclone Megh originated from the Arabian Sea causing even more rains in parts of Bari region in Puntland and Somaliland. The storm produced a maximum windspeed of 110knots.

Areas affected included: Af Kalahay, Alula, Bareda, BiyoCade, Boolimoog, Dhurbo, Fagoora, Geesalay, Murcanyo, Sayn Weyn, Sayn Yar, Toxin and Xaabo.

Re-estimated population figures after Tropical Megh, showed 4.9 million people were in need of assistance, 308,700 children under-5 were acutely malnourished, of which 55,800 were severely malnourished and 1.1 million remain in a protracted internal displacement situation.

Each row in this dataset represents geolocated point observations. Data was last updated on HDX on 2026-04-05. Geographic scope: SOM.

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

---

Dataset Characteristics

Domain	Forced displacement and migration
Unit of observation	Geolocated point observations
Rows (total)	36
Columns	32 (21 numeric, 11 categorical, 0 datetime)
Train split	28 rows
Test split	7 rows
Geographic scope	SOM
Publisher	IGAD Climate Prediction and Applications Center (ICPAC)
HDX last updated	2026-04-05

---

Variables

Geographic — cy (range 5.0–5.0), yyyymmddhh (range 2015110518.0–2015111012.0), lat (range 12.2–13.7), long (range 46.7–63.2), vmax (range 20.0–110.0) and 3 others.

Identifier / Metadata — fid (a__2015_TC_Megh.1, a__2015_TC_Megh.2, a__2015_TC_Megh.21), fid_1 (range 1.0–36.0), windcode (NEQ), stormname (MEGH), esa_source (HDX) and 1 others.

Other — the_geom (POINT (54.9 12.6), POINT (50.8 12.2), POINT (58.3 12.8)), basin (IO), tech (BEST), tau (range 0.0–0.0), mslp (range 941.0–1007.0) and 13 others.

---

Quick Start

from datasets import load_dataset

ds    = load_dataset("electricsheepafrica/africa-icpac-geonode-2015-tropical-cyclone-megh-path")
train = ds["train"].to_pandas()
test  = ds["test"].to_pandas()

print(train.shape)
train.head()

---

Schema

Column	Type	Null %	Range / Sample Values
fid	object	0.0%	a__2015_TC_Megh.1, a__2015_TC_Megh.2, a__2015_TC_Megh.21
fid_1	int64	0.0%	1.0 – 36.0 (mean 18.5)
the_geom	object	0.0%	POINT (54.9 12.6), POINT (50.8 12.2), POINT (58.3 12.8)
basin	object	0.0%	IO
cy	int64	0.0%	5.0 – 5.0 (mean 5.0)
yyyymmddhh	int64	0.0%	2015110518.0 – 2015111012.0 (mean 2015110789.5556)
tech	object	0.0%	BEST
tau	int64	0.0%	0.0 – 0.0 (mean 0.0)
lat	float64	0.0%	12.2 – 13.7 (mean 12.725)
long	float64	0.0%	46.7 – 63.2 (mean 54.8417)
vmax	int64	0.0%	20.0 – 110.0 (mean 71.5278)
mslp	int64	0.0%	941.0 – 1007.0 (mean 969.3056)
ty	object	0.0%	TY, TS, TD
rad	int64	0.0%	0.0 – 0.0 (mean 0.0)
windcode	object	8.3%	NEQ
rad1	int64	0.0%	0.0 – 60.0 (mean 33.1944)
rad2	int64	0.0%	0.0 – 55.0 (mean 31.3889)
rad3	int64	0.0%	0.0 – 60.0 (mean 31.3889)
rad4	int64	0.0%	0.0 – 60.0 (mean 33.4722)
radp	int64	0.0%	1008.0 – 1011.0 (mean 1009.0833)
rrp	int64	0.0%	120.0 – 200.0 (mean 150.8333)
mrd	int64	0.0%	5.0 – 40.0 (mean 16.6944)
gusts	int64	0.0%	0.0 – 0.0 (mean 0.0)
eye	int64	0.0%	0.0 – 0.0 (mean 0.0)
subregn	object	5.6%	A
maxseas	int64	0.0%	0.0 – 0.0 (mean 0.0)
dir	int64	0.0%	0.0 – 0.0 (mean 0.0)
speed	int64	0.0%
stormname	object	0.0%	MEGH
depth	object	5.6%	D, M
esa_source	object	0.0%	HDX
esa_processed	object	0.0%

---

Numeric Summary

Column	Min	Max	Mean	Median
fid_1	1.0	36.0	18.5	18.5
cy	5.0	5.0	5.0	5.0
yyyymmddhh	2015110518.0	2015111012.0	2015110789.5556	2015110806.0
tau	0.0	0.0	0.0	0.0
lat	12.2	13.7	12.725	12.7
long	46.7	63.2	54.8417	54.9
vmax	20.0	110.0	71.5278	75.0
mslp	941.0	1007.0	969.3056	966.5
rad	0.0	0.0	0.0	0.0
rad1	0.0	60.0	33.1944	30.0
rad2	0.0	55.0	31.3889	30.0
rad3	0.0	60.0	31.3889	30.0
rad4	0.0	60.0	33.4722	30.0
radp	1008.0	1011.0	1009.0833	1009.0
rrp	120.0	200.0	150.8333	155.0

---

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. 8 column(s) with >80% missing values were removed: technum, initials, seas, seascode, seas1, seas2.... 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 IGAD Climate Prediction and Applications Center (ICPAC) and has not been independently validated by ESA.
• Automated cleaning cannot correct for misreported values, definitional inconsistencies, or sampling bias in the original collection.
• Refer to the original HDX dataset page for the publisher's own methodology notes and caveats.

---

Citation

@dataset{hdx_africa_icpac_geonode_2015_tropical_cyclone_megh_path,
  title     = {2015 Tropical Cyclone Megh Path},
  author    = {IGAD Climate Prediction and Applications Center (ICPAC)},
  year      = {2026},
  url       = {https://data.humdata.org/dataset/icpac-geonode-2015-tropical-cyclone-megh-path},
  note      = {Repackaged for machine learning by Electric Sheep Africa (https://huggingface.co/electricsheepafrica)}
}

---

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