---
language:
- sw
- en
license: other
license_name: world-bank-microdata-terms
license_link: https://microdata.worldbank.org/index.php/terms-of-use
tags:
- agriculture
- financial-inclusion
- survey-data
- smallholder-farmers
- africa
- tanzania
- tabular
pretty_name: "CGAP Smallholder Survey - Tanzania 2015"
size_categories:
- 1K<n<10K
task_categories:
- tabular-classification
- tabular-regression
---

# CGAP Smallholder Household Survey - Tanzania (2015)

## Dataset Description

Nationally representative household survey of smallholder farming families in **Tanzania**, conducted by the Consultative Group to Assist the Poor (CGAP) in 2015.

### Dataset Summary

| Metric | Value |
|--------|-------|
| **Country** | Tanzania |
| **Year** | 2015 |
| **Households Surveyed** | ~3,000 |
| **Total Records** | 10,822 |

### Tables

| Table | Rows | Columns | Description |
|-------|------|---------|-------------|
| `household` | 2,993 | 275 | Household-level characteristics |
| `single` | 2,795 | 1,017 | Single respondent detailed survey |
| `multiple` | 5,034 | 420 | Repeated measures (members, plots, livestock) |

## Usage

```python
from datasets import load_dataset

# Load household-level data
household = load_dataset("YOUR_USERNAME/cgap-smallholder-tza", "household", split="train")

# Load single respondent data
single = load_dataset("YOUR_USERNAME/cgap-smallholder-tza", "single", split="train")

# Load multiple respondent data
multiple = load_dataset("YOUR_USERNAME/cgap-smallholder-tza", "multiple", split="train")
```

### With Pandas

```python
import pandas as pd
df = pd.read_parquet("hf://datasets/YOUR_USERNAME/cgap-smallholder-tza/household.parquet")
```

## Features

The dataset covers:

- **Demographics**: Household composition, age, gender, education levels
- **Agriculture**: Crops cultivated, land holdings, livestock, farming practices
- **Financial Services**: Bank accounts, mobile money usage, savings groups, credit access
- **Income & Expenditure**: Income sources, consumption patterns
- **Assets**: Household and productive assets
- **Shocks**: Economic shocks and coping mechanisms


## Documentation

### Official CGAP User Guide

For detailed variable definitions, survey methodology, and questionnaire documentation, see the official CGAP user guide:

📖 **[Tanzania User Guide](https://www.cgap.org/sites/default/files/working_paper_cgap_smallholder_survey_user_guide_tanzania_may_2016.pdf)**

### Data Dictionary

This repository includes comprehensive data dictionaries with statistics for every variable:

- **[DATA_DICTIONARY.md](DATA_DICTIONARY.md)** - Human-readable markdown format
- **[data_dictionary.json](data_dictionary.json)** - Machine-readable JSON format

The data dictionary includes for each variable:
- Data type (numeric/categorical)
- Non-null counts and missing value percentages
- Value ranges (min, max, mean, median) for numeric columns
- Unique value counts and distributions for categorical columns

## Suggested ML Tasks

- Financial inclusion prediction (binary classification)
- Income/wealth classification
- Household segmentation (clustering)
- Agricultural productivity prediction
- Vulnerability/poverty risk assessment

## Source

- **Publisher**: Consultative Group to Assist the Poor (CGAP)
- **Host**: World Bank Microdata Library
- **URL**: https://www.cgap.org/research/data/smallholder-families-data-hub

## Citation

```bibtex
@misc{cgap_smallholder_tza_2015,
  title={CGAP Smallholder Families Household Survey - Tanzania},
  author={{Consultative Group to Assist the Poor}},
  year={2015},
  publisher={World Bank Microdata Library}
}
```

## License

Subject to World Bank Microdata Library [Terms of Use](https://microdata.worldbank.org/index.php/terms-of-use). Intended for research and educational purposes.