electricsheepafrica/africa-unsom-sgrhumanrights

--- annotations_creators: - no-annotation language_creators: - found language: - en license: cc-by-4.0 multilinguality: - monolingual size_categories: - n<1K source_datasets: - original task_categories: - tabular-classification task_ids: [] tags: - africa - humanitarian - hdx - electric-sheep-africa - complex-emergency-conflict-security - gender - peacekeeping - som pretty_name: "Peace and Security Pillar: Human Rights Situation in Somalia, Data From Secretary General Reports" dataset_info: splits: - name: train num_examples: 21 - name: test num_examples: 5 --- # Peace and Security Pillar: Human Rights Situation in Somalia, Data From Secretary General Reports **Publisher:** United Nations Peace and Security Data Hub · **Source:** [HDX](https://data.humdata.org/dataset/unsom-sgrhumanrights) · **License:** `cc-by-igo` · **Updated:** 2026-04-09 --- ## Abstract This dataset was last updated in April 2023 and will no longer receive updates. Historical data remains available for reference. This dataset is an extraction of information from the Secretary General's Report on the situation in Somalia (SG reports) since 2013. It provides quarterly total of civilian casualties, fatalities, persons arbitrarily arrested, children subject to grave violations, to reflect the country's human rights situation. Each row in this dataset represents discrete events or incidents. Temporal coverage is indicated by the `start_of_reporting_period`, `end_of_reporting_period` column(s). Geographic scope: **SOM**. *Curated into ML-ready Parquet format by [Electric Sheep Africa](https://huggingface.co/electricsheepafrica).* --- ## Dataset Characteristics | | | |---|---| | **Domain** | Conflict and security | | **Unit of observation** | Discrete events or incidents | | **Rows (total)** | 27 | | **Columns** | 23 (12 numeric, 8 categorical, 3 datetime) | | **Train split** | 21 rows | | **Test split** | 5 rows | | **Geographic scope** | SOM | | **Publisher** | United Nations Peace and Security Data Hub | | **HDX last updated** | 2026-04-09 | --- ## Variables **Geographic** — `country_or_area` (range 706.0–706.0), `number_of_persons_arbitrarily_arrested` (range 2.0–27.0), `breakdown_of_persons_arbitrarily_arrested` (Targeted attacks continued on journalists and human rights defenders, in particular on those defending victims in cases involving gender-based violence. Ten journalists (three in Kismaayo, three in Mogadishu and four in “Somaliland”) were arrested. No specific measures were taken to reinforce the protection of journalists, notwithstanding a statement by the Federal Government of Somalia indicating its intention to investigate attacks on journalists and bring the perpetrators to justice., In total, 66 individuals were arrested for issues related to the rights to freedom of expression and of assembly, including three journalists, one civil society activist in Puntland and one clan elder in “Somaliland”. Three television stations and one newspaper outlet were banned, while one of the television stations remains under suspension in “Somaliland”. On a positive note, the licenses of two media outlets that had been previously revoked were reinstated in “Somaliland” and a female poet who had been sentenced to three years’ imprisonment in Hargeisa was released following a presidential pardon., Several incidents illustrated continued challenges to freedom of expression. A journalist was sentenced to 18 months in prison for defamation and false news in “Somaliland”, while two journalists were arrested and released without ch arges after reporting on a demonstration in Jubaland. The state of emergency imposed in Galmudug entailed restrictions on freedom of expression and information. A journalist died following injuries sustained from an Al -Shabaab attack in Beledweyne, in the Hiran region, in September. In “Somaliland”, one media outlet was closed, reportedly for broadcasting false news, and access to social media was blocked during the presidential elections in November.), `number_of_children_subject_to_grave_violations` (range 58.0–1654.0), `number_of_male_children_subject_to_grave_violations` (range 50.0–1372.0) and 3 others. **Temporal** — `start_of_reporting_period`, `end_of_reporting_period`, `last_update`. **Demographic** — `number_of_female_victims_of_sgbv` (range 2.0–158.0), `number_of_male_victims_of_sgbv` (range 0.0–5.0). **Outcome / Measurement** — `number_of_civilian_casualties` (range 123.0–1301.0), `number_of_civilian_fatalities` (range 82.0–382.0), `number_of_children_victims_of_sgbv` (range 11.0–96.0). **Identifier / Metadata** — `id` (range 1.0–27.0), `m49_code` (SOM), `source` (S/2013/709, S/2019/393, S/2022/665), `esa_source` (HDX), `esa_processed` (2026-04-09). **Other** — `breakdown_of_civilian_casualties` (During the reporting period, a total of 123 civilian casualties were documented, compared with 113 over the previous reporting period. Of those, Al -Shabaab or unidentified armed individuals were implicated in the killing of 95 people, with Somali security forces implicated in the killing of 27 people and AMISOM forces implicated in 22 killings., There was a rise in human rights violations during the reporting period, due primarily to security operations and Al-Shabaab activity. Reported civilian casualties totalled 387, including 159 deaths and 228 injuries. Operations by security forces resulted in 83 civilian casualties, of which 16 deaths and 31 injuries were attributed to the Somali security forces; 5 deaths and 9 injuries to AMISOM; and 9 deaths, including of 4 children, and 11 injuries to air strikes by the Kenyan military operating bilaterally in Gedo region. Local elders denounced reported air strikes and later engaged the Government of Kenya, resulting in an agreement to cooperate to prevent civilian casualties. No outcome has been reported so far from the investigative committee established by the office of the Prime Minister of Somalia after the air strikes by the Kenyan military in Gedo in January, which reportedly resulted in 29 deaths and the destruction of livestock, water wells and houses., During the reporting period, UNSOM recorded 419 Somali civilian casualties (173 killed and 246 injured), a 2 per cent decrease compared with 428 civilian casualties recorded during the previous period. Unidentified actors were the main perpetrators, responsible for 249 civilian casualties (59 per cent), followed by 88 (21 per cent) attributed to Al-Shabaab, 46 (11 per cent) to state security forces, 30 (7 per cent) to clan militias and 6 (1 per cent) to the Liyu Police.), `breakdown_of_sgbv` (Gender-based violence remains a key protection concern, as over 80 per cent of recorded incidents have affected internally displaced persons, and 96 per cent of the survivors of such violence are women and girls. As at 15 August, 40 cases of conflict related sexual violence had been reported during the reporting period, involving elements of the Somali National Army, regional security forces, clan militias and Al-Shabaab., The United Nations continued to receive reports of sexual violence from across the country, mostly affecting displaced women and girls living in settlements for internally displaced persons. The monitoring and reporting mechanism, through its analysis reporting arrangements, received six reported cases of conflict-related sexual violence in September and October. Fifty-nine incidents of rape and other sexual violence, involving 58 girls and one boy, were reported by the country task force on monitoring and reporting., In total, from January to April, 56 incidents affecting 62 victims/survivors of conflict-related sexual violence were documented. All those affected were women and girls (56 minors and 6 adults).). --- ## Quick Start ```python from datasets import load_dataset ds = load_dataset("electricsheepafrica/africa-unsom-sgrhumanrights") 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% | 1.0 – 27.0 (mean 14.0) | | `country_or_area` | int64 | 0.0% | 706.0 – 706.0 (mean 706.0) | | `m49_code` | object | 0.0% | SOM | | `start_of_reporting_period` | datetime64[ns] | 0.0% | | | `end_of_reporting_period` | datetime64[ns] | 0.0% | | | `number_of_civilian_casualties` | float64 | 25.9% | 123.0 – 1301.0 (mean 464.2) | | `number_of_civilian_fatalities` | float64 | 51.9% | 82.0 – 382.0 (mean 180.4615) | | `breakdown_of_civilian_casualties` | object | 25.9% | During the reporting period, a total of 123 civilian casualties were documented, compared with 113 over the previous reporting period. Of those, Al -Shabaab or unidentified armed individuals were implicated in the killing of 95 people, with Somali security forces implicated in the killing of 27 people and AMISOM forces implicated in 22 killings., There was a rise in human rights violations during the reporting period, due primarily to security operations and Al-Shabaab activity. Reported civilian casualties totalled 387, including 159 deaths and 228 injuries. Operations by security forces resulted in 83 civilian casualties, of which 16 deaths and 31 injuries were attributed to the Somali security forces; 5 deaths and 9 injuries to AMISOM; and 9 deaths, including of 4 children, and 11 injuries to air strikes by the Kenyan military operating bilaterally in Gedo region. Local elders denounced reported air strikes and later engaged the Government of Kenya, resulting in an agreement to cooperate to prevent civilian casualties. No outcome has been reported so far from the investigative committee established by the office of the Prime Minister of Somalia after the air strikes by the Kenyan military in Gedo in January, which reportedly resulted in 29 deaths and the destruction of livestock, water wells and houses., During the reporting period, UNSOM recorded 419 Somali civilian casualties (173 killed and 246 injured), a 2 per cent decrease compared with 428 civilian casualties recorded during the previous period. Unidentified actors were the main perpetrators, responsible for 249 civilian casualties (59 per cent), followed by 88 (21 per cent) attributed to Al-Shabaab, 46 (11 per cent) to state security forces, 30 (7 per cent) to clan militias and 6 (1 per cent) to the Liyu Police. | | `number_of_persons_arbitrarily_arrested` | float64 | 44.4% | 2.0 – 27.0 (mean 9.9333) | | `breakdown_of_persons_arbitrarily_arrested` | object | 44.4% | Targeted attacks continued on journalists and human rights defenders, in particular on those defending victims in cases involving gender-based violence. Ten journalists (three in Kismaayo, three in Mogadishu and four in “Somaliland”) were arrested. No specific measures were taken to reinforce the protection of journalists, notwithstanding a statement by the Federal Government of Somalia indicating its intention to investigate attacks on journalists and bring the perpetrators to justice., In total, 66 individuals were arrested for issues related to the rights to freedom of expression and of assembly, including three journalists, one civil society activist in Puntland and one clan elder in “Somaliland”. Three television stations and one newspaper outlet were banned, while one of the television stations remains under suspension in “Somaliland”. On a positive note, the licenses of two media outlets that had been previously revoked were reinstated in “Somaliland” and a female poet who had been sentenced to three years’ imprisonment in Hargeisa was released following a presidential pardon., Several incidents illustrated continued challenges to freedom of expression. A journalist was sentenced to 18 months in prison for defamation and false news in “Somaliland”, while two journalists were arrested and released without ch arges after reporting on a demonstration in Jubaland. The state of emergency imposed in Galmudug entailed restrictions on freedom of expression and information. A journalist died following injuries sustained from an Al -Shabaab attack in Beledweyne, in the Hiran region, in September. In “Somaliland”, one media outlet was closed, reportedly for broadcasting false news, and access to social media was blocked during the presidential elections in November. | | `number_of_children_subject_to_grave_violations` | float64 | 3.7% | 58.0 – 1654.0 (mean 630.8846) | | `number_of_male_children_subject_to_grave_violations` | float64 | 3.7% | 50.0 – 1372.0 (mean 513.9231) | | `number_of_female_children_subject_to_grave_violations` | float64 | 3.7% | 8.0 – 282.0 (mean 115.8462) | | `breakdown_of_grave_violations_against_children` | object | 3.7% | During the reporting period, 449 cases of grave violations against children (355 boys and 65 girls), in particular abduction, recruitment and killing and maiming of children, were documented. Most incidents were reported from the southern and central regions of Somalia. Since the beginning of 2013, 1,100 children (863 boys and 237 girls) formerly associated with armed forces or groups and children with other vulnerabilities (including those in street situations), benefited from reintegration programmes. Of those children, 110 (65 boys and 45 girls) are enrolled in formal schools, while the rest are in vocational training., A total of 61 incidents of grave violations affecting 58 children (50 boys and 8 girls) were reported. The violations mostly involved the recruitment and use of children by security forces linked to the army and all ied militias, the police and Al-Shabaab, sexual violence and arbitrary arrest and detention. The standard operating procedures for handling children formerly associated with armed forces and groups were signed by the Minister of Defence and the Minister of National Security., From 7 to 31 May, the country task force on monitoring and reporting on grave violations against children in armed conflict verified 153 grave violations against 151 children (135 boys and 16 girls). A total of 72 children were victims of killing and maiming, 21 were abducted and 17 were subjected to rape and other forms of sexual violence. Attacks on two schools were also verified. Of the violations, 71 (46 per cent) were attributed to Al-Shabaab, 44 (29 per cent) to the federal and state armed forces, 32 (21 per cent) to unknown armed elements, 3 (2 per cent) to clan militia and 3 (2 per cent) to Ahl al-Sunna wal-Jama‘a. | | `number_of_incidents_of_conflict_related_sexual_and_gender_based_violence` | float64 | 59.3% | 2.0 – 163.0 (mean 30.5455) | | `number_of_female_victims_of_sgbv` | float64 | 66.7% | 2.0 – 158.0 (mean 33.1111) | | `number_of_male_victims_of_sgbv` | float64 | 70.4% | 0.0 – 5.0 (mean 0.75) | | `number_of_children_victims_of_sgbv` | float64 | 44.4% | 11.0 – 96.0 (mean 60.8) | | `breakdown_of_sgbv` | object | 59.3% | Gender-based violence remains a key protection concern, as over 80 per cent of recorded incidents have affected internally displaced persons, and 96 per cent of the survivors of such violence are women and girls. As at 15 August, 40 cases of conflict related sexual violence had been reported during the reporting period, involving elements of the Somali National Army, regional security forces, clan militias and Al-Shabaab., The United Nations continued to receive reports of sexual violence from across the country, mostly affecting displaced women and girls living in settlements for internally displaced persons. The monitoring and reporting mechanism, through its analysis reporting arrangements, received six reported cases of conflict-related sexual violence in September and October. Fifty-nine incidents of rape and other sexual violence, involving 58 girls and one boy, were reported by the country task force on monitoring and reporting., In total, from January to April, 56 incidents affecting 62 victims/survivors of conflict-related sexual violence were documented. All those affected were women and girls (56 minors and 6 adults). | | `source` | object | 0.0% | S/2013/709, S/2019/393, S/2022/665 | | `last_update` | datetime64[ns] | 0.0% | | | `esa_source` | object | 0.0% | HDX | | `esa_processed` | object | 0.0% | 2026-04-09 | --- ## Numeric Summary | Column | Min | Max | Mean | Median | |---|---|---|---|---| | `id` | 1.0 | 27.0 | 14.0 | 14.0 | | `country_or_area` | 706.0 | 706.0 | 706.0 | 706.0 | | `number_of_civilian_casualties` | 123.0 | 1301.0 | 464.2 | 380.0 | | `number_of_civilian_fatalities` | 82.0 | 382.0 | 180.4615 | 159.0 | | `number_of_persons_arbitrarily_arrested` | 2.0 | 27.0 | 9.9333 | 6.0 | | `number_of_children_subject_to_grave_violations` | 58.0 | 1654.0 | 630.8846 | 582.5 | | `number_of_male_children_subject_to_grave_violations` | 50.0 | 1372.0 | 513.9231 | 437.5 | | `number_of_female_children_subject_to_grave_violations` | 8.0 | 282.0 | 115.8462 | 122.0 | | `number_of_incidents_of_conflict_related_sexual_and_gender_based_violence` | 2.0 | 163.0 | 30.5455 | 8.0 | | `number_of_female_victims_of_sgbv` | 2.0 | 158.0 | 33.1111 | 8.0 | | `number_of_male_victims_of_sgbv` | 0.0 | 5.0 | 0.75 | 0.0 | | `number_of_children_victims_of_sgbv` | 11.0 | 96.0 | 60.8 | 68.0 | --- ## 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`. 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 United Nations Peace and Security Data Hub 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: `number_of_civilian_casualties`, `number_of_civilian_fatalities`, `breakdown_of_civilian_casualties`, `number_of_persons_arbitrarily_arrested`, `breakdown_of_persons_arbitrarily_arrested`, `number_of_incidents_of_conflict_related_sexual_and_gender_based_violence`, `number_of_female_victims_of_sgbv`, `number_of_male_victims_of_sgbv`.... - Refer to the [original HDX dataset page](https://data.humdata.org/dataset/unsom-sgrhumanrights) for the publisher's own methodology notes and caveats. --- ## Citation ```bibtex @dataset{hdx_africa_unsom_sgrhumanrights, title = {Peace and Security Pillar: Human Rights Situation in Somalia, Data From Secretary General Reports}, author = {United Nations Peace and Security Data Hub}, year = {2026}, url = {https://data.humdata.org/dataset/unsom-sgrhumanrights}, 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: - 无标注 language_creators: - 现有公开资源 language: - 英语 license: CC-BY-4.0 multilinguality: - 单语言 size_categories: - 样本量<1000 source_datasets: - 原创数据集 task_categories: - 表格分类 task_ids: [] tags: - 非洲 - 人道主义 - HDX - Electric Sheep Africa - 复杂紧急冲突安全 - 性别 - 维和 - 索马里（SOM） pretty_name: "和平与安全支柱：索马里人权局势——来自联合国秘书长报告的数据" dataset_info: splits: - name: 训练集 num_examples: 21 - name: 测试集 num_examples: 5 --- # 和平与安全支柱：索马里人权局势——来自联合国秘书长报告的数据 **发布方**：联合国和平与安全数据中心 · **来源**：[HDX（人道主义数据交换平台，Humanitarian Data Exchange）](https://data.humdata.org/dataset/unsom-sgrhumanrights) · **授权协议**：`cc-by-igo` · **更新时间**：2026-04-09 --- ## 摘要 本数据集最后一次更新于2023年4月，此后将不再进行更新，历史数据仍可查阅参考。 本数据集提取自2013年以来联合国秘书长关于索马里局势的报告（以下简称秘书长报告），按季度统计平民伤亡、死亡人数、任意逮捕人数以及遭受严重侵害的儿童数量，以反映该国的人权状况。 数据集中的每一行代表一起独立事件或事故。数据的时间覆盖范围由`start_of_reporting_period（报告起始周期）`和`end_of_reporting_period（报告结束周期）`列标注。地理覆盖范围：**索马里（SOM）**。 *本数据集经[Electric Sheep Africa](https://huggingface.co/electricsheepafrica)整理为适合机器学习使用的Parquet格式。* --- ## 数据集特征 | | | |---|---| | **应用领域** | 冲突与安全 | | **观测单元** | 独立事件或事故 | | **总行数** | 27 | | **总列数** | 23（12个数值列、8个分类列、3个日期时间列） | | **训练集划分** | 21条数据 | | **测试集划分** | 5条数据 | | **地理覆盖范围** | 索马里（SOM） | | **发布方** | 联合国和平与安全数据中心 | | **HDX平台最后更新时间** | 2026-04-09 | --- ## 变量说明 **地理类变量**：`country_or_area`（取值范围706.0–706.0）、`number_of_persons_arbitrarily_arrested`（任意逮捕人数，取值范围2.0–27.0）、`breakdown_of_persons_arbitrarily_arrested`（任意逮捕情况说明，涵盖记者与人权捍卫者袭击、媒体管制等事件）、`number_of_children_subject_to_grave_violations`（遭受严重侵害的儿童数量，取值范围58.0–1654.0）、`number_of_male_children_subject_to_grave_violations`（男性遭受严重侵害儿童数量，取值范围50.0–1372.0）以及其余3个地理类变量。 **时间类变量**：`start_of_reporting_period`（报告起始周期）、`end_of_reporting_period`（报告结束周期）、`last_update`（最后更新时间）。 **人口统计类变量**：`number_of_female_victims_of_sgbv`（性与基于性别的暴力（Sexual and Gender-Based Violence, SGBV）女性受害者数量，取值范围2.0–158.0）、`number_of_male_victims_of_sgbv`（性与基于性别的暴力男性受害者数量，取值范围0.0–5.0）。 **结果/衡量类变量**：`number_of_civilian_casualties`（平民伤亡总人数，取值范围123.0–1301.0）、`number_of_civilian_fatalities`（平民死亡人数，取值范围82.0–382.0）、`number_of_children_victims_of_sgbv`（性与基于性别的暴力儿童受害者数量，取值范围11.0–96.0）。 **标识符/元数据类变量**：`id`（数据ID，取值范围1.0–27.0）、`m49_code`（M49国家代码，值为SOM）、`source`（报告来源，包括S/2013/709、S/2019/393、S/2022/665）、`esa_source`（数据来源平台，HDX）、`esa_processed`（数据处理时间，2026-04-09）。 **其他类变量**：`breakdown_of_civilian_casualties`（平民伤亡情况说明）、`breakdown_of_sgbv`（性与基于性别的暴力情况说明）。 --- ## 快速入门 python from datasets import load_dataset ds = load_dataset("electricsheepafrica/africa-unsom-sgrhumanrights") train = ds["train"].to_pandas() test = ds["test"].to_pandas() print(train.shape) train.head() --- ## 数据结构 | 列名 | 数据类型 | 缺失率 | 取值范围/示例值 | |---|---|---|---| | `id` | int64 | 0.0% | 1.0 – 27.0（均值14.0） | | `country_or_area` | int64 | 0.0% | 706.0 – 706.0（均值706.0） | | `m49_code` | object | 0.0% | SOM | | `start_of_reporting_period` | datetime64[ns] | 0.0% | 无 | | `end_of_reporting_period` | datetime64[ns] | 0.0% | 无 | | `number_of_civilian_casualties` | float64 | 25.9% | 123.0 – 1301.0（均值464.2） | | `number_of_civilian_fatalities` | float64 | 51.9% | 82.0 – 382.0（均值180.4615） | | `breakdown_of_civilian_casualties` | object | 25.9% | 长文本说明，涵盖本报告期平民伤亡统计、肇事者归属等详细信息 | | `number_of_persons_arbitrarily_arrested` | float64 | 44.4% | 2.0 – 27.0（均值9.9333） | | `breakdown_of_persons_arbitrarily_arrested` | object | 44.4% | 长文本说明，涵盖记者逮捕、媒体管制等相关事件 | | `number_of_children_subject_to_grave_violations` | float64 | 3.7% | 58.0 – 1654.0（均值630.8846） | | `number_of_male_children_subject_to_grave_violations` | float64 | 3.7% | 50.0 – 1372.0（均值513.9231） | | `number_of_female_children_subject_to_grave_violations` | float64 | 3.7% | 8.0 – 282.0（均值115.8462） | | `breakdown_of_grave_violations_against_children` | object | 3.7% | 长文本说明，涵盖儿童严重侵害事件统计、肇事者归属等 | | `number_of_incidents_of_conflict_related_sexual_and_gender_based_violence` | float64 | 59.3% | 2.0 – 163.0（均值30.5455） | | `number_of_female_victims_of_sgbv` | float64 | 66.7% | 2.0 – 158.0（均值33.1111） | | `number_of_male_victims_of_sgbv` | float64 | 70.4% | 0.0 – 5.0（均值0.75） | | `number_of_children_victims_of_sgbv` | float64 | 44.4% | 11.0 – 96.0（均值60.8） | | `breakdown_of_sgbv` | object | 59.3% | 长文本说明，涵盖性与基于性别的暴力事件统计、受害者情况等 | | `source` | object | 0.0% | S/2013/709, S/2019/393, S/2022/665 | | `last_update` | datetime64[ns] | 0.0% | 无 | | `esa_source` | object | 0.0% | HDX | | `esa_processed` | object | 0.0% | 2026-04-09 | --- ## 数值型变量统计摘要 | 列名 | 最小值 | 最大值 | 均值 | 中位数 | |---|---|---|---|---| | `id` | 1.0 | 27.0 | 14.0 | 14.0 | | `country_or_area` | 706.0 | 706.0 | 706.0 | 706.0 | | `number_of_civilian_casualties` | 123.0 | 1301.0 | 464.2 | 380.0 | | `number_of_civilian_fatalities` | 82.0 | 382.0 | 180.4615 | 159.0 | | `number_of_persons_arbitrarily_arrested` | 2.0 | 27.0 | 9.9333 | 6.0 | | `number_of_children_subject_to_grave_violations` | 58.0 | 1654.0 | 630.8846 | 582.5 | | `number_of_male_children_subject_to_grave_violations` | 50.0 | 1372.0 | 513.9231 | 437.5 | | `number_of_female_children_subject_to_grave_violations` | 8.0 | 282.0 | 115.8462 | 122.0 | | `number_of_incidents_of_conflict_related_sexual_and_gender_based_violence` | 2.0 | 163.0 | 30.5455 | 8.0 | | `number_of_female_victims_of_sgbv` | 2.0 | 158.0 | 33.1111 | 8.0 | | `number_of_male_victims_of_sgbv` | 0.0 | 5.0 | 0.75 | 0.0 | | `number_of_children_victims_of_sgbv` | 11.0 | 96.0 | 60.8 | 68.0 | --- ## 数据整理流程 原始数据通过CKAN 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（snake_case）。将常见缺失值标记（`N/A`、`null`、`none`、`-`、`unknown`、`no data`、`#N/A`）统一替换为`NaN`。根据解析成功率（阈值>85%），将3列从字符串类型转换为数值型或日期时间型。数据集以固定随机种子（42）按80/20的比例划分为训练集和测试集，并保存为Snappy压缩的Parquet格式。 --- ## 数据集局限性 - 数据来源于联合国和平与安全数据中心，并未经Electric Sheep Africa独立验证。 - 自动化清洗无法修正原始数据收集过程中存在的错报、定义不一致或抽样偏差问题。 - 以下列的缺失率超过20%，在建模时需谨慎使用：`number_of_civilian_casualties`、`number_of_civilian_fatalities`、`breakdown_of_civilian_casualties`、`number_of_persons_arbitrarily_arrested`、`breakdown_of_persons_arbitrarily_arrested`、`number_of_incidents_of_conflict_related_sexual_and_gender_based_violence`、`number_of_female_victims_of_sgbv`、`number_of_male_victims_of_sgbv`等。 - 请参阅[原始HDX数据集页面](https://data.humdata.org/dataset/unsom-sgrhumanrights)查看发布方提供的方法说明和注意事项。 --- ## 引用格式 bibtex @dataset{hdx_africa_unsom_sgrhumanrights, title = {和平与安全支柱：索马里人权局势——来自联合国秘书长报告的数据}, author = {联合国和平与安全数据中心}, year = {2026}, url = {https://data.humdata.org/dataset/unsom-sgrhumanrights}, note = {由Electric Sheep Africa（https://huggingface.co/electricsheepafrica）重新打包为机器学习可用格式} } --- *[Electric Sheep Africa](https://huggingface.co/electricsheepafrica) — 非洲机器学习数据集基础设施，尼日利亚拉各斯。*