electricsheepafrica/africa-learning-levels-in-kenya-2015

---
annotations_creators:
- no-annotation
language_creators:
- found
language:
- en
license: cc-by-4.0
multilinguality:
- monolingual
size_categories:
- 100K<n<1M
source_datasets:
- original
task_categories:
- tabular-regression
- other
task_ids: []
tags:
- africa
- humanitarian
- hdx
- electric-sheep-africa
- ken
pretty_name: "Learning levels in Kenya (2015)"
dataset_info:
splits:
- name: train
num_examples: 153035
- name: test
num_examples: 38258
---

# Learning levels in Kenya (2015)

**Publisher:** Uwezo at Twaweza East Africa (inactive) · **Source:** [HDX](https://data.humdata.org/dataset/learning-levels-in-kenya-2015) · **License:** `cc-by` · **Updated:** 2025-02-19

---

## Abstract

These datasets contains data relevant to learner achievement in Kenya. This ranges from school enrollment, socioeconomic indicators, school infrastructure among others.

Each row in this dataset represents subnational administrative unit observations. Data was last updated on HDX on 2025-02-19. Geographic scope: **KEN**.

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

---

## Dataset Characteristics

| | |
|---|---|
| **Domain** | Water, sanitation and hygiene (wash) |
| **Unit of observation** | Subnational administrative unit observations |
| **Rows (total)** | 191,294 |
| **Columns** | 112 (81 numeric, 31 categorical, 0 datetime) |
| **Train split** | 153,035 rows |
| **Test split** | 38,258 rows |
| **Geographic scope** | KEN |
| **Publisher** | Uwezo at Twaweza East Africa (inactive) |
| **HDX last updated** | 2025-02-19 |

---

## Variables

**Geographic** — `id_district` (range 101.0–820.0), `id_districtname` (Wajir East, Wajir South, Mandera Central), `id_region` (range 1.0–8.0), `id_regionname` (Rift Valley, Eastern, Western), `location` (TOWNSHIP, MARALAL TOWN, MALINDI TOWN) and 15 others.

**Demographic** — `id_hh` (range 1004.0–4700000.0), `id_village` (range 1.0–4740.0), `census_hholds` (range 30.0–1599.0), `hhno` (range 1.0–119.0), `hhlistno` (range 1.0–200.0) and 17 others.

**Identifier / Metadata** — `id_database` (KE15), `division_name` (CENTRAL, MUKOGODO, TINDERET), `eacode` (range 100000000000.0–820000000000.0), `water_source` (range 1.0–7.0), `prova_name` and 2 others.

**Other** — `use_tele` (range 1.0–2.0), `visited` (range 1.0–2.0), `house_wall`, `house_lighting`, `water_treatment` (range 1.0–6.0) and 58 others.

---

## Quick Start

```python
from datasets import load_dataset

ds = load_dataset("electricsheepafrica/africa-learning-levels-in-kenya-2015")
train = ds["train"].to_pandas()
test = ds["test"].to_pandas()

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

---

## Schema

| Column | Type | Null % | Range / Sample Values |
|---|---|---|---|
| `id_database` | object | 0.0% | KE15 |
| `id_district` | int64 | 0.0% | 101.0 – 820.0 (mean 576.1508) |
| `id_districtname` | object | 0.0% | Wajir East, Wajir South, Mandera Central |
| `id_hh` | float64 | 0.0% | 1004.0 – 4700000.0 (mean 2565862.8367) |
| `id_region` | int64 | 0.0% | 1.0 – 8.0 (mean 5.624) |
| `id_regionname` | object | 0.0% | Rift Valley, Eastern, Western |
| `id_village` | int64 | 0.0% | 1.0 – 4740.0 (mean 2565.0841) |
| `division_name` | object | 0.0% | CENTRAL, MUKOGODO, TINDERET |
| `eacode` | float64 | 0.0% | 100000000000.0 – 820000000000.0 (mean 576562882264.9952) |
| `location` | object | 0.0% | TOWNSHIP, MARALAL TOWN, MALINDI TOWN |
| `sub_location` | object | 0.0% | TOWNSHIP, MAJENGO, LANGAS |
| `census_hholds` | int64 | 0.0% | 30.0 – 1599.0 (mean 112.528) |
| `village_estate` | object | 0.0% | BONDENI, KIMUGUL, CENTRAL |
| `ea_type` | object | 0.0% | Rural, Urban, 11 |
| `hhno` | int64 | 0.0% | 1.0 – 119.0 (mean 9.9719) |
| `hhlistno` | float64 | 3.4% | 1.0 – 200.0 (mean 32.8422) |
| `person_answering` | object | 5.0% | Mother, Father, Guardian |
| `hh_gender` | object | 4.8% | Male, Female, 38 |
| `hh_age` | float64 | 4.0% | 0.0 – 110.0 (mean 41.3431) |
| `hh_edu_raw` | float64 | 4.1% | 1.0 – 5.0 (mean 3.9529) |
| `use_tele` | float64 | 27.1% | 1.0 – 2.0 (mean 1.1145) |
| `visited` | float64 | 8.4% | 1.0 – 2.0 (mean 1.8915) |
| `home_language` | object | 1.9% | |
| `q101_male` | float64 | 3.4% | 1.0 – 25.0 (mean 3.1252) |
| `q101_female` | float64 | 1.9% | 1.0 – 25.0 (mean 3.2256) |
| `hh_size` | float64 | 1.2% | 1.0 – 47.0 (mean 6.2666) |
| `house_wall` | object | 1.2% | |
| `house_lighting` | object | 0.0% | |
| `water_source` | float64 | 4.6% | 1.0 – 7.0 (mean 3.6754) |
| `water_treatment` | float64 | 5.8% | 1.0 – 6.0 (mean 2.8998) |
| `mealsperday` | float64 | 4.4% | 1.0 – 3.0 (mean 2.7747) |
| `eat_veg` | int64 | 0.0% | 0.0 – 1.0 (mean 0.9778) |
| `eat_fruit` | int64 | 0.0% | 0.0 – 1.0 (mean 0.9293) |
| `q108_toilet` | float64 | 2.7% | |
| `asset_tv` | int64 | 0.0% | |
| `asset_radio` | int64 | 0.0% | |
| `asset_computer` | int64 | 0.0% | |
| `asset_phone` | int64 | 0.0% | |
| `asset_car` | int64 | 0.0% | |
| `asset_cattle` | int64 | 0.0% | |
| `asset_donkey` | int64 | 0.0% | |
| `asset_camel` | int64 | 0.0% | |
| `asset_sheep_goat` | int64 | 0.0% | |
| `asset_bicycle` | int64 | 0.0% | |
| `asset_motorbike` | int64 | 0.0% | |
| `asset_cart` | float64 | 0.0% | |
| `q110_popular_radio` | object | 32.4% | |
| `q110_popular_presenter` | object | 41.0% | |
| `q110_popular_tv_station` | object | 72.6% | |
| `q110_favourite_newspaper` | object | 75.4% | |
| `q112` | float64 | 73.0% | |
| `administer_feedback` | float64 | 11.6% | |
| `uwezocommu` | float64 | 12.1% | |
| `childno` | int64 | 0.0% | |
| `age` | float64 | 0.0% | |
| `gender` | object | 0.0% | |
| `disability` | float64 | 1.3% | |
| `mothers_toschool` | float64 | 10.5% | |
| `mothers_edu_raw` | float64 | 43.0% | |
| `mothers_age` | float64 | 13.9% | |
| `preschool` | float64 | 0.0% | |
| `grade` | float64 | 44.4% | |
| `schooltype` | object | 43.6% | |
| `schoolmatch` | float64 | 0.0% | |
| `neverenrolled` | float64 | 24.6% | |
| `dropout` | float64 | 24.6% | |
| `missyesterday` | float64 | 36.5% | |
| `testsample` | float64 | 61.6% | |
| `english` | object | 36.0% | |
| `english1` | float64 | 64.6% | |
| `english2` | float64 | 64.7% | |
| `swahili` | object | 37.2% | |
| `swahili1` | float64 | 62.7% | |
| `swahili2` | float64 | 62.8% | |
| `math` | object | 37.0% | |
| `ethno1` | float64 | 39.2% | |
| `ethno2` | float64 | 39.6% | |
| `bonus1` | float64 | 36.8% | |
| `bonus2` | float64 | 37.3% | |
| `bonus3` | float64 | 37.2% | |
| `muac_reading` | float64 | 69.8% | |
| `asset_elec` | float64 | 0.0% | |
| `hh_edu` | object | 30.8% | |
| `asset_water` | float64 | 4.6% | |
| `mothers_edu` | object | 43.1% | |
| `completed` | float64 | 35.8% | |
| `enr_ans` | float64 | 24.6% | |
| `hh_children` | float64 | 0.0% | |
| `weight` | float64 | 0.0% | |
| `asset_raw` | float64 | 11.8% | |
| `asset_wealth_index` | float64 | 11.8% | |
| `quin_asset_wealth` | float64 | 11.8% | |
| `tri_asset_wealth` | float64 | 11.8% | |
| `prova_name` | object | 0.0% | |
| `district_nohhlds` | float64 | 0.0% | |
| `noschooldata` | int64 | 0.0% | |
| `nohhldsinea` | int64 | 0.0% | |
| `noeasindistrict` | int64 | 0.0% | |
| `nodistricts` | float64 | 0.0% | |
| `english_imputed` | object | 31.4% | |
| `english1_imputed` | float64 | 31.4% | |
| `english2_imputed` | float64 | 31.4% | |
| `swahili_imputed` | object | 31.4% | |
| `swahili1_imputed` | float64 | 31.4% | |
| `swahili2_imputed` | float64 | 31.4% | |
| `math_imputed` | object | 31.5% | |
| `bonus1_imputed` | float64 | 31.5% | |
| `bonus2_imputed` | float64 | 31.5% | |
| `bonus3_imputed` | float64 | 31.5% | |
| `county` | object | 0.0% | |
| `esa_source` | object | 0.0% | |
| `esa_processed` | object | 0.0% | |

---

## Numeric Summary

| Column | Min | Max | Mean | Median |
|---|---|---|---|---|
| `id_district` | 101.0 | 820.0 | 576.1508 | 614.0 |
| `id_hh` | 1004.0 | 4700000.0 | 2565862.8367 | 2600000.0 |
| `id_region` | 1.0 | 8.0 | 5.624 | 6.0 |
| `id_village` | 1.0 | 4740.0 | 2565.0841 | 2598.0 |
| `eacode` | 100000000000.0 | 820000000000.0 | 576562882264.9952 | 610000000000.0 |
| `census_hholds` | 30.0 | 1599.0 | 112.528 | 99.0 |
| `hhno` | 1.0 | 119.0 | 9.9719 | 10.0 |
| `hhlistno` | 1.0 | 200.0 | 32.8422 | 21.0 |
| `hh_age` | 0.0 | 110.0 | 41.3431 | 39.0 |
| `hh_edu_raw` | 1.0 | 5.0 | 3.9529 | 4.0 |
| `use_tele` | 1.0 | 2.0 | 1.1145 | 1.0 |
| `visited` | 1.0 | 2.0 | 1.8915 | 2.0 |
| `q101_male` | 1.0 | 25.0 | 3.1252 | 3.0 |
| `q101_female` | 1.0 | 25.0 | 3.2256 | 3.0 |
| `hh_size` | 1.0 | 47.0 | 6.2666 | 6.0 |

---

## 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`. 10 column(s) with >80% missing values were removed: `q111gvt`, `q111pvt`, `q111other`, `q111home`, `dropout_year`, `dropout_class`.... 16 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 Uwezo at Twaweza East Africa (inactive) 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: `use_tele`, `q110_popular_radio`, `q110_popular_presenter`, `q110_popular_tv_station`, `q110_favourite_newspaper`, `q112`, `mothers_edu_raw`, `grade`....
- Refer to the [original HDX dataset page](https://data.humdata.org/dataset/learning-levels-in-kenya-2015) for the publisher's own methodology notes and caveats.

---

## Citation

```bibtex
@dataset{hdx_africa_learning_levels_in_kenya_2015,
title = {Learning levels in Kenya (2015)},
author = {Uwezo at Twaweza East Africa (inactive)},
year = {2025},
url = {https://data.humdata.org/dataset/learning-levels-in-kenya-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.*