electricsheepafrica/africa-drc-displacement-forecasts

---
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
- tabular-regression
task_ids: []
tags:
- africa
- humanitarian
- hdx
- electric-sheep-africa
- asylum-seekers
- displacement
- forecasting
- internally-displaced-persons-idp
- refugees
- afg
- bfa
- bdi
- cmr
- caf
pretty_name: "Future Displacement Forecasts"
dataset_info:
splits:
- name: train
num_examples: 20
- name: test
num_examples: 5
---

# Future Displacement Forecasts

**Publisher:** Danish Refugee Council · **Source:** [HDX](https://data.humdata.org/dataset/drc-displacement-forecasts) · **License:** `cc-by` · **Updated:** 2026-02-22

---

## Abstract

Forecasts of forced displacement (IDPs, asylum seekers and refugees) one to three years into the future based on machine learning model.

Each row in this dataset represents country-level aggregates. Data was last updated on HDX on 2026-02-22. Geographic scope: **AFG, BFA, BDI, CMR, CAF, TCD, COL, COD, and 18 others**.

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

---

## Dataset Characteristics

| | |
|---|---|
| **Domain** | Forced displacement and migration |
| **Unit of observation** | Country-level aggregates |
| **Rows (total)** | 26 |
| **Columns** | 13 (9 numeric, 4 categorical, 0 datetime) |
| **Train split** | 20 rows |
| **Test split** | 5 rows |
| **Geographic scope** | AFG, BFA, BDI, CMR, CAF, TCD, COL, COD, and 18 others |
| **Publisher** | Danish Refugee Council |
| **HDX last updated** | 2026-02-22 |

---

## Variables

**Geographic** — `country_name` (Afghanistan, Myanmar, Burundi), `country_code` (AFG, MMR, BDI).

**Identifier / Metadata** — `esa_source` (HDX), `esa_processed` (2026-04-07).

**Other** — `2015` (range -0.9146–0.5809), `2016` (range -0.8845–0.7106), `2017` (range -57500000000.0–947000000000.0), `2018` (range -0.7326–0.8803), `2019` (range -0.8209–0.2595) and 4 others.

---

## Quick Start

```python
from datasets import load_dataset

ds = load_dataset("electricsheepafrica/africa-drc-displacement-forecasts")
train = ds["train"].to_pandas()
test = ds["test"].to_pandas()

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

---

## Schema

| Column | Type | Null % | Range / Sample Values |
|---|---|---|---|
| `country_name` | object | 0.0% | Afghanistan, Myanmar, Burundi |
| `country_code` | object | 0.0% | AFG, MMR, BDI |
| `2015` | float64 | 23.1% | -0.9146 – 0.5809 (mean -0.1645) |
| `2016` | float64 | 0.0% | -0.8845 – 0.7106 (mean -0.0236) |
| `2017` | float64 | 0.0% | -57500000000.0 – 947000000000.0 (mean 34211538461.4929) |
| `2018` | float64 | 0.0% | -0.7326 – 0.8803 (mean -0.0908) |
| `2019` | float64 | 0.0% | -0.8209 – 0.2595 (mean -0.116) |
| `2020` | float64 | 3.8% | -0.5663 – 0.2766 (mean -0.0508) |
| `2021` | float64 | 3.8% | -0.3136 – 0.4539 (mean 0.0041) |
| `2022` | float64 | 0.0% | -0.9192 – 0.4765 (mean -0.0141) |
| `2023` | float64 | 0.0% | -0.59 – 1.1355 (mean 0.0684) |
| `esa_source` | object | 0.0% | HDX |
| `esa_processed` | object | 0.0% | 2026-04-07 |

---

## Numeric Summary

| Column | Min | Max | Mean | Median |
|---|---|---|---|---|
| `2015` | -0.9146 | 0.5809 | -0.1645 | -0.0711 |
| `2016` | -0.8845 | 0.7106 | -0.0236 | -0.0053 |
| `2017` | -57500000000.0 | 947000000000.0 | 34211538461.4929 | -0.0107 |
| `2018` | -0.7326 | 0.8803 | -0.0908 | -0.0225 |
| `2019` | -0.8209 | 0.2595 | -0.116 | -0.0775 |
| `2020` | -0.5663 | 0.2766 | -0.0508 | -0.007 |
| `2021` | -0.3136 | 0.4539 | 0.0041 | -0.0439 |
| `2022` | -0.9192 | 0.4765 | -0.0141 | 0.037 |
| `2023` | -0.59 | 1.1355 | 0.0684 | 0.0066 |

---

## 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) with >80% missing values were removed: `2010`, `2011`, `2012`, `2013`, `2014`. 1 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 Danish Refugee Council 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: `2015`.
- This dataset spans 26 countries; geographic and methodological inconsistencies across national boundaries may affect cross-country comparability.
- Refer to the [original HDX dataset page](https://data.humdata.org/dataset/drc-displacement-forecasts) for the publisher's own methodology notes and caveats.

---

## Citation

```bibtex
@dataset{hdx_africa_drc_displacement_forecasts,
title = {Future Displacement Forecasts},
author = {Danish Refugee Council},
year = {2026},
url = {https://data.humdata.org/dataset/drc-displacement-forecasts},
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.*