electricsheepafrica/africa-somalia-3w-q1-2015

---
annotations_creators:
- no-annotation
language_creators:
- found
language:
- en
license: cc-by-4.0
multilinguality:
- monolingual
size_categories:
- 1K<n<10K
source_datasets:
- original
task_categories:
- tabular-classification
- other
task_ids: []
tags:
- africa
- humanitarian
- hdx
- electric-sheep-africa
- education
- food-security
- health
- nutrition
- shelter
- water-sanitation-and-hygiene-wash
- who-is-doing-what-and-where-3w-4w-5w
- som
pretty_name: "Somalia - Who is doing what and where (3W) - 2015"
dataset_info:
splits:
- name: train
num_examples: 3137
- name: test
num_examples: 784
---

# Somalia - Who is doing what and where (3W) - 2015

**Publisher:** OCHA Somalia · **Source:** [HDX](https://data.humdata.org/dataset/somalia-3w-q1-2015) · **License:** `cc-by-igo` · **Updated:** 2023-11-15

---

## Abstract

This is 3W data for Somalia from clusters (WASH, Health, Nutrition, Shelter, Food Security, Education and Protection)

Each row in this dataset represents subnational administrative unit observations. Temporal coverage is indicated by the `start_date`, `end_date` column(s). Geographic scope: **SOM**.

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

---

## Dataset Characteristics

| | |
|---|---|
| **Domain** | Food security and nutrition |
| **Unit of observation** | Subnational administrative unit observations |
| **Rows (total)** | 3,922 |
| **Columns** | 14 (1 numeric, 11 categorical, 2 datetime) |
| **Train split** | 3,137 rows |
| **Test split** | 784 rows |
| **Geographic scope** | SOM |
| **Publisher** | OCHA Somalia |
| **HDX last updated** | 2023-11-15 |

---

## Variables

**Geographic** — `activity_description` (Response activities to GBV (case management, PEP kits, medical assistance, legal assistance for the survivors, safe house) , Teacher Incentives, Learning Materials), `region` (Banadir, Gedo, Bay), `district` (Baidoa, Jowhar, Hargeysa).

**Temporal** — `start_date`, `end_date`.

**Identifier / Metadata** — `esa_source` (HDX), `esa_processed`.

**Other** — `sector` (Nutrition, Protection , WASH), `organisation` (WFP, UNICEF, INGO), `implementing_partner` (Undisclosed, WVI, MOH SL), `sub_sector` (TSFP, OTP, Gender-Based Violence), `status` (Open, Ongoing, Completed) and 2 others.

---

## Quick Start

```python
from datasets import load_dataset

ds = load_dataset("electricsheepafrica/africa-somalia-3w-q1-2015")
train = ds["train"].to_pandas()
test = ds["test"].to_pandas()

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

---

## Schema

| Column | Type | Null % | Range / Sample Values |
|---|---|---|---|
| `sector` | object | 0.0% | Nutrition, Protection , WASH |
| `organisation` | object | 32.7% | WFP, UNICEF, INGO |
| `implementing_partner` | object | 0.0% | Undisclosed, WVI, MOH SL |
| `sub_sector` | object | 9.0% | TSFP, OTP, Gender-Based Violence |
| `activity_description` | object | 9.0% | Response activities to GBV (case management, PEP kits, medical assistance, legal assistance for the survivors, safe house) , Teacher Incentives, Learning Materials |
| `region` | object | 0.6% | Banadir, Gedo, Bay |
| `district` | object | 2.1% | Baidoa, Jowhar, Hargeysa |
| `status` | object | 23.3% | Open, Ongoing, Completed |
| `no_of_beneficiaries` | float64 | 68.2% | 0.0 – 500000.0 (mean 5133.8796) |
| `unit` | object | 64.4% | IND, Ind, PLANS |
| `start_date` | datetime64[ns] | 34.4% | |
| `end_date` | datetime64[ns] | 51.5% | |
| `esa_source` | object | 0.0% | HDX |
| `esa_processed` | object | 0.0% | |

---

## Numeric Summary

| Column | Min | Max | Mean | Median |
|---|---|---|---|---|
| `no_of_beneficiaries` | 0.0 | 500000.0 | 5133.8796 | 180.0 |

---

## 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`. 1,737 exact duplicate rows were removed. 3 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 OCHA Somalia 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: `organisation`, `status`, `no_of_beneficiaries`, `unit`, `start_date`, `end_date`.
- Refer to the [original HDX dataset page](https://data.humdata.org/dataset/somalia-3w-q1-2015) for the publisher's own methodology notes and caveats.

---

## Citation

```bibtex
@dataset{hdx_africa_somalia_3w_q1_2015,
title = {Somalia - Who is doing what and where (3W) - 2015},
author = {OCHA Somalia},
year = {2023},
url = {https://data.humdata.org/dataset/somalia-3w-q1-2015},
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.*