electricsheepafrica/africa-electricaldistributiongridmaps

---
annotations_creators:
- no-annotation
language_creators:
- found
language:
- en
license: cc-by-4.0
multilinguality:
- monolingual
size_categories:
- 10K<n<100K
source_datasets:
- original
task_categories:
- other
task_ids: []
tags:
- africa
- humanitarian
- hdx
- electric-sheep-africa
- facilities-infrastructure
- geodata
- civ
- cod
- mwi
- nga
- uga
pretty_name: "Electrical Distribution Grid Maps"
dataset_info:
splits:
- name: train
num_examples: 18284
- name: test
num_examples: 4571
---

# Electrical Distribution Grid Maps

**Publisher:** AI for Good at Meta · **Source:** [HDX](https://data.humdata.org/dataset/electricaldistributiongridmaps) · **License:** `cc-by` · **Updated:** 2025-11-19

---

## Abstract

Facebook has produced a model to help map global medium voltage (MV) grid infrastructure, i.e. the distribution lines which connect high-voltage transmission infrastructure to consumer-serving low-voltage distribution. The data found here are model outputs for six select African countries: Malawi, Nigeria, Uganda, DRC, Cote D’Ivoire, and Zambia. The grid maps are produced using a new methodology that employs various publicly-available datasets (night time satellite imagery, roads, political boundaries, etc) to predict the location of existing MV grid infrastructure. The model documentation and code are also available , so data scientists and planners globally can replicate the model to expand model coverage to other countries where this data is not already available. You can find the model code and documentation here: https://github.com/facebookresearch/many-to-many-dijkstra

Note: current model accuracy is approximately 70% when compared to existing ground-truthed data. Accuracy can be further improved by integrating other locally-relevant information into the model and running it again.

Resolution: geotiff is provided at Bing Tile Level 20

Each row in this dataset represents geolocated point observations. Data was last updated on HDX on 2025-11-19. Geographic scope: **CIV, COD, MWI, NGA, UGA, ZMB**.

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

---

## Dataset Characteristics

| | |
|---|---|
| **Domain** | Humanitarian and development data |
| **Unit of observation** | Geolocated point observations |
| **Rows (total)** | 22,856 |
| **Columns** | 5 (3 numeric, 2 categorical, 0 datetime) |
| **Train split** | 18,284 rows |
| **Test split** | 4,571 rows |
| **Geographic scope** | CIV, COD, MWI, NGA, UGA, ZMB |
| **Publisher** | AI for Good at Meta |
| **HDX last updated** | 2025-11-19 |

---

## Variables

**Geographic** — `lat` (range 4.4067–10.6228), `lon` (range -8.5089–-2.5873).

**Identifier / Metadata** — `electrical_distribution_grid` (range 1.0–1.0), `esa_source` (HDX), `esa_processed` (2026-04-04).

---

## Quick Start

```python
from datasets import load_dataset

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

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

---

## Schema

| Column | Type | Null % | Range / Sample Values |
|---|---|---|---|
| `lat` | float64 | 0.0% | 4.4067 – 10.6228 (mean 7.2465) |
| `lon` | float64 | 0.0% | -8.5089 – -2.5873 (mean -5.4782) |
| `electrical_distribution_grid` | int64 | 0.0% | 1.0 – 1.0 (mean 1.0) |
| `esa_source` | object | 0.0% | HDX |
| `esa_processed` | object | 0.0% | 2026-04-04 |

---

## Numeric Summary

| Column | Min | Max | Mean | Median |
|---|---|---|---|---|
| `lat` | 4.4067 | 10.6228 | 7.2465 | 7.0518 |
| `lon` | -8.5089 | -2.5873 | -5.4782 | -5.4547 |
| `electrical_distribution_grid` | 1.0 | 1.0 | 1.0 | 1.0 |

---

## 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`. 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 AI for Good at Meta and has not been independently validated by ESA.
- Automated cleaning cannot correct for misreported values, definitional inconsistencies, or sampling bias in the original collection.
- This dataset spans 6 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/electricaldistributiongridmaps) for the publisher's own methodology notes and caveats.

---

## Citation

```bibtex
@dataset{hdx_africa_electricaldistributiongridmaps,
title = {Electrical Distribution Grid Maps},
author = {AI for Good at Meta},
year = {2025},
url = {https://data.humdata.org/dataset/electricaldistributiongridmaps},
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.*