electricsheepafrica/africa-somalia-operational-presence

---
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:
- other
task_ids: []
tags:
- africa
- humanitarian
- hdx
- electric-sheep-africa
- hxl
- operational-presence
- who-is-doing-what-and-where-3w-4w-5w
- som
pretty_name: "Somalia: Operational Presence"
dataset_info:
splits:
- name: train
num_examples: 7747
- name: test
num_examples: 1936
---

# Somalia: Operational Presence

**Publisher:** OCHA Somalia · **Source:** [HDX](https://data.humdata.org/dataset/somalia-operational-presence) · **License:** `cc-by` · **Updated:** 2026-04-03

---

## Abstract

The Who does What Where (3W) is a core humanitarian coordination dataset. It is critical to know where humanitarian organizations are working, what they are doing and their capability in order to identify gaps, avoid duplication of efforts, and plan for future humanitarian response (if needed). The data includes a list of humanitarian organizations by district and cluster, as well as a unique count of organizations.

Each row in this dataset represents subnational administrative unit observations. Data was last updated on HDX on 2026-04-03. Geographic scope: **SOM**.

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

---

## Dataset Characteristics

| | |
|---|---|
| **Domain** | Humanitarian and development data |
| **Unit of observation** | Subnational administrative unit observations |
| **Rows (total)** | 9,684 |
| **Columns** | 24 (5 numeric, 19 categorical, 0 datetime) |
| **Train split** | 7,747 rows |
| **Test split** | 1,936 rows |
| **Geographic scope** | SOM |
| **Publisher** | OCHA Somalia |
| **HDX last updated** | 2026-04-03 |

---

## Variables

**Geographic** — `activity_description` (Raising awareness on the risks/threat posed explosive hazards, MRE sessions, Relief: General Food Distribution (GFD)), `region` (Bay, Mudug, Banadir), `district` (Baidoa, Gaalkacyo, Belet Weyne), `specific_location` (Baidoa, Kismayo, Belet Weyne), `start_date_dd_mm_yy` and 4 others.

**Demographic** — `total_of_individuals` (range 0.0–317122.7774), `of_households`.

**Identifier / Metadata** — `programme_name` (ICSP, TSFP, OTP), `esa_source`, `esa_processed`.

**Other** — `cluster` (Protection, Education, WASH), `organization` (WFP, RI, SCI), `implementing_partner` (RI, HALO Trust, SCI), `donor` (OFDA, OCHA, WFP), `rural_urban` (Rural, Urban, IDPs/Non IDPs) and 5 others.

---

## Quick Start

```python
from datasets import load_dataset

ds = load_dataset("electricsheepafrica/africa-somalia-operational-presence")
train = ds["train"].to_pandas()
test = ds["test"].to_pandas()

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

---

## Schema

| Column | Type | Null % | Range / Sample Values |
|---|---|---|---|
| `cluster` | object | 0.0% | Protection, Education, WASH |
| `organization` | object | 0.0% | WFP, RI, SCI |
| `implementing_partner` | object | 0.0% | RI, HALO Trust, SCI |
| `programme_name` | object | 61.7% | ICSP, TSFP, OTP |
| `activity_description` | object | 16.5% | Raising awareness on the risks/threat posed explosive hazards, MRE sessions, Relief: General Food Distribution (GFD) |
| `donor` | object | 34.4% | OFDA, OCHA, WFP |
| `region` | object | 0.0% | Bay, Mudug, Banadir |
| `district` | object | 0.0% | Baidoa, Gaalkacyo, Belet Weyne |
| `specific_location` | object | 2.0% | Baidoa, Kismayo, Belet Weyne |
| `rural_urban` | object | 62.1% | Rural, Urban, IDPs/Non IDPs |
| `status` | object | 0.1% | |
| `duration` | object | 65.3% | |
| `start_date_dd_mm_yy` | object | 1.9% | |
| `end_date_dd_mm_yy` | object | 1.4% | |
| `y_coord` | object | 63.2% | |
| `x_coord` | object | 63.1% | |
| `total_of_individuals` | float64 | 20.4% | 0.0 – 317122.7774 (mean 1316.2217) |
| `of_households` | object | 65.3% | |
| `women` | float64 | 33.1% | 0.0 – 103579.63 (mean 486.0282) |
| `men` | float64 | 35.6% | 0.0 – 107807.37 (mean 330.0472) |
| `girls` | float64 | 43.4% | 0.0 – 50985.0 (mean 419.6685) |
| `boys` | float64 | 32.6% | 0.0 – 54981.7587 (mean 10269.4661) |
| `esa_source` | object | 0.0% | |
| `esa_processed` | object | 0.0% | |

---

## Numeric Summary

| Column | Min | Max | Mean | Median |
|---|---|---|---|---|
| `total_of_individuals` | 0.0 | 317122.7774 | 1316.2217 | 64.0 |
| `women` | 0.0 | 103579.63 | 486.0282 | 31.0 |
| `men` | 0.0 | 107807.37 | 330.0472 | 17.0 |
| `girls` | 0.0 | 50985.0 | 419.6685 | 30.0 |
| `boys` | 0.0 | 54981.7587 | 10269.4661 | 80.0 |

---

## 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`. 9 column(s) with >80% missing values were removed: `modality`, `conditionality`, `restriction`, `transfer_value_usd`, `frequency`, `delivery_mechanism`.... 6,055 exact duplicate rows were removed. 4 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: `programme_name`, `donor`, `rural_urban`, `duration`, `y_coord`, `x_coord`, `total_of_individuals`, `of_households`....
- Refer to the [original HDX dataset page](https://data.humdata.org/dataset/somalia-operational-presence) for the publisher's own methodology notes and caveats.

---

## Citation

```bibtex
@dataset{hdx_africa_somalia_operational_presence,
title = {Somalia: Operational Presence},
author = {OCHA Somalia},
year = {2026},
url = {https://data.humdata.org/dataset/somalia-operational-presence},
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.*

annotations_creators:
- 无注释
language_creators:
- 现有数据源
language:
- 英语
license: 知识共享署名4.0国际许可协议（CC BY 4.0）
multilinguality:
- 单语言
size_categories:
- 1000<样本数<10000
source_datasets:
- 原始数据集
task_categories:
- 其他
task_ids: []
tags:
- 非洲
- 人道主义
- HDX（Humanitarian Data Exchange）
- Electric Sheep Africa
- HXL
- 运营态势
- 谁在何地开展行动（3W/4W/5W）
- 索马里（SOM）
pretty_name: "索马里：运营态势"
dataset_info:
splits:
- name: train
num_examples: 7747
- name: test
num_examples: 1936

# 索马里：运营态势

**发布方：** 联合国人道主义事务协调厅索马里办事处（OCHA Somalia） · **来源：** [HDX（Humanitarian Data Exchange）](https://data.humdata.org/dataset/somalia-operational-presence) · **许可协议：** `CC BY` · **最后更新：** 2026-04-03

---

## 摘要

**谁在何地开展行动（3W/4W/5W）** 是核心的人道主义协调数据集。明确人道主义组织的工作地点、开展的活动及其能力，对于识别协作缺口、避免工作重复、规划未来人道主义响应（如有需要）至关重要。本数据集包含按地区和行动集群划分的人道主义组织清单，以及组织的唯一计数。本数据集的每一行均代表次国家级行政单元的观测数据。本数据集最后更新于HDX平台的时间为2026年4月3日。地理覆盖范围：**索马里（SOM）**。

*本数据集已由[Electric Sheep Africa](https://huggingface.co/electricsheepafrica)整理为适用于机器学习的Parquet格式。*

---

## 数据集特征

| | |
|---|---|
| **领域** | 人道主义与发展数据 |
| **观测单元** | 次国家级行政单元观测样本 |
| **总样本量** | 9684条 |
| **列数** | 24列（5个数值型，19个分类型，0个日期时间型） |
| **训练集样本量** | 7747条 |
| **测试集样本量** | 1936条 |
| **地理覆盖范围** | 索马里（SOM） |
| **发布方** | 联合国人道主义事务协调厅索马里办事处 |
| **HDX最后更新时间** | 2026年4月3日 |

---

## 变量说明

**地理类变量**：`activity_description`（如“提升爆炸物风险/威胁认知、MRE培训、救济：通用粮食分配（GFD）”）、`region`（如巴伊州、穆杜格州、班纳迪尔州）、`district`（如拜多阿、加尔卡尤、贝莱特韦内）、`specific_location`（如拜多阿、基斯马尤、贝莱特韦内）、`start_date_dd_mm_yy`及另外4个变量。
**人口统计类变量**：`total_of_individuals`（取值范围0.0–317122.7774）、`of_households`。
**标识符/元数据类变量**：`programme_name`（如ICSP、TSFP、OTP）、`esa_source`、`esa_processed`。
**其他类变量**：`cluster`（如保护、教育、WASH（水、卫生与个人卫生））、`organization`（如WFP、RI、SCI）、`implementing_partner`（如RI、HALO Trust、SCI）、`donor`（如OFDA、OCHA、WFP）、`rural_urban`（如农村、城市、境内流离失所者/非境内流离失所者）及另外5个变量。

---

## 快速入门

python
from datasets import load_dataset

ds = load_dataset("electricsheepafrica/africa-somalia-operational-presence")
train = ds["train"].to_pandas()
test = ds["test"].to_pandas()

print(train.shape)
train.head()

---

## 数据模式

| 列名 | 数据类型 | 缺失率 | 取值范围/示例值 |
|---|---|---|---|
| `cluster` | 字符型 | 0.0% | 保护、教育、WASH（水、卫生与个人卫生） |
| `organization` | 字符型 | 0.0% | WFP、RI、SCI |
| `implementing_partner` | 字符型 | 0.0% | RI、HALO Trust、SCI |
| `programme_name` | 字符型 | 61.7% | ICSP、TSFP、OTP |
| `activity_description` | 字符型 | 16.5% | 提升爆炸物风险/威胁认知、MRE培训、救济：通用粮食分配（GFD） |
| `donor` | 字符型 | 34.4% | OFDA、OCHA、WFP |
| `region` | 字符型 | 0.0% | 巴伊州、穆杜格州、班纳迪尔州 |
| `district` | 字符型 | 0.0% | 拜多阿、加尔卡尤、贝莱特韦内 |
| `specific_location` | 字符型 | 2.0% | 拜多阿、基斯马尤、贝莱特韦内 |
| `rural_urban` | 字符型 | 62.1% | 农村、城市、境内流离失所者/非境内流离失所者 |
| `status` | 字符型 | 0.1% | 无 |
| `duration` | 字符型 | 65.3% | 无 |
| `start_date_dd_mm_yy` | 字符型 | 1.9% | 无 |
| `end_date_dd_mm_yy` | 字符型 | 1.4% | 无 |
| `y_coord` | 字符型 | 63.2% | 无 |
| `x_coord` | 字符型 | 63.1% | 无 |
| `total_of_individuals` | 浮点型 | 20.4% | 0.0 – 317122.7774（均值1316.2217） |
| `of_households` | 字符型 | 65.3% | 无 |
| `women` | 浮点型 | 33.1% | 0.0 – 103579.63（均值486.0282） |
| `men` | 浮点型 | 35.6% | 0.0 – 107807.37（均值330.0472） |
| `girls` | 浮点型 | 43.4% | 0.0 – 50985.0（均值419.6685） |
| `boys` | 浮点型 | 32.6% | 0.0 – 54981.7587（均值10269.4661） |
| `esa_source` | 字符型 | 0.0% | 无 |
| `esa_processed` | 字符型 | 0.0% | 无 |

---

## 数值型变量统计摘要

| 列名 | 最小值 | 最大值 | 均值 | 中位数 |
|---|---|---|---|---|
| `total_of_individuals` | 0.0 | 317122.7774 | 1316.2217 | 64.0 |
| `women` | 0.0 | 103579.63 | 486.0282 | 31.0 |
| `men` | 0.0 | 107807.37 | 330.0472 | 17.0 |
| `girls` | 0.0 | 50985.0 | 419.6685 | 30.0 |
| `boys` | 0.0 | 54981.7587 | 10269.4661 | 80.0 |

---

## 数据整理流程

原始数据通过CKAN应用程序编程接口（API）从HDX平台下载，并转换为Parquet格式。列名统一转换为小写并标准化为蛇形命名法。将常见的缺失值标记（`N/A`、`null`、`none`、`-`、`unknown`、`no data`、`#N/A`）统一替换为`NaN`。移除了9个缺失率超过80%的列：`modality`、`conditionality`、`restriction`、`transfer_value_usd`、`frequency`、`delivery_mechanism`等。移除了6055条完全重复的样本。根据解析成功率（阈值>85%），将4个列从字符型转换为数值型或日期时间型。本数据集以固定随机种子（42）按80/20的比例划分为训练集与测试集，并保存为Snappy压缩的Parquet格式。

---

## 局限性

- 本数据集源自联合国人道主义事务协调厅索马里办事处，未由Electric Sheep Africa进行独立验证。
- 自动化清洗流程无法修正原始数据收集中的错报值、定义不一致或抽样偏差问题。
- 以下列的缺失率超过20%，在建模过程中需谨慎使用：`programme_name`、`donor`、`rural_urban`、`duration`、`y_coord`、`x_coord`、`total_of_individuals`、`of_households`等。
- 请参阅[原始HDX数据集页面](https://data.humdata.org/dataset/somalia-operational-presence)获取发布方的方法说明与免责声明。

---

## 引用

bibtex
@dataset{hdx_africa_somalia_operational_presence,
title = {索马里：运营态势},
author = {联合国人道主义事务协调厅索马里办事处},
year = {2026},
url = {https://data.humdata.org/dataset/somalia-operational-presence},
note = {由Electric Sheep Africa重新打包为机器学习可用数据集（https://huggingface.co/electricsheepafrica)}
}

---

*[Electric Sheep Africa](https://huggingface.co/electricsheepafrica) — 非洲机器学习数据集基础设施提供商，尼日利亚拉各斯。*