electricsheepafrica/climate-adaptation-tech-ssa-synthetic

---
license: cc-by-4.0
task_categories:
- tabular-regression
- tabular-classification
tags:
- agriculture
- food-security
- africa
- synthetic-data
- climate-adaptation
- technology-adoption
- smallholder-farming
- climate-change
- resilience
size_categories:
- 1M<n<10M
language:
- en
---

# Climate Adaptation & Technology - Sub-Saharan Africa Synthetic Dataset

## Dataset Summary

**Dataset 4 in the African Agriculture Portfolio**

This dataset contains **1,000,000 synthetic household records** focused on climate adaptation strategies and technology adoption in Sub-Saharan African agriculture. Building on Datasets 1-3 (farm systems, livestock, and post-harvest), this adds critical climate change response variables.

**Key Features:**
- 🌍 **57 variables** (41 inherited + 16 new climate/tech variables)
- 🌡️ **Climate exposure**: Droughts, floods, temperature trends
- 🚜 **Technology adoption**: Seeds, irrigation, conservation, mechanization, digital tools
- 🏗️ **Adaptation strategies**: Diversification, migration responses
- 💰 **Support access**: Credit, insurance, climate finance
- 📚 **16 literature sources** for new variables
- 🔒 **100% synthetic**: No real households
- ✅ **Literature-grounded**: All parameters from peer-reviewed research

## Dataset Details

### Dataset Description

**Size**: 1,000,000 rows × 57 variables
**Format**: CSV (617 MB) + Parquet (193 MB)
**License**: CC-BY-4.0
**Created**: November 2025
**Version**: 1.0
**Random Seed**: 42 (reproducible)

### New Variables (16 Climate & Technology)

#### **Climate Shocks & Exposure (4 variables)**

| Variable | Type | Description | Range |
|----------|------|-------------|-------|
| `drought_events_5yr` | count | Droughts in past 5 years | 0-5 events |
| `flood_exposure` | binary | Experienced flood damage | yes/no |
| `temperature_trend` | continuous | Perceived temp change (°C) | -1 to +3°C |
| `climate_info_access` | binary | Access to weather forecasts | yes/no |

**Key Stats**:
- Average **1.8 droughts per 5 years** (matches Sheffield et al. 2014)
- **22% flood exposure** rate (matches regional data)
- **27% have climate info access** (growing with mobile penetration)

#### **Technology Adoption (6 variables)**

| Variable | Type | Description | Adoption Rate |
|----------|------|-------------|---------------|
| `improved_seed_use` | binary | Hybrid/improved seeds | ~18% |
| `irrigation_access` | categorical | Irrigation type | 7% total |
| `conservation_agriculture` | binary | CA practices | ~12% |
| `agroforestry` | binary | Tree integration | ~30% |
| `mechanization_level` | categorical | Land prep tech | 60% manual |
| `digital_ag_tools` | binary | Smartphone ag apps | ~10% |

**Reflects reality**: Low tech adoption rates typical of SSA smallholders

#### **Adaptation Strategies (3 variables)**

| Variable | Type | Description |
|----------|------|-------------|
| `crop_diversification_index` | continuous | Crop diversity (0-100) |
| `livelihood_diversification` | categorical | Off-farm income levels |
| `migration_response` | binary | Climate-induced migration |

**Key Stats**:
- Mean crop diversity: **48.7** (moderate diversification)
- **60% have some off-farm income** (coping strategy)
- **15% climate-induced migration** (vulnerable areas)

#### **Support Access (3 variables)**

| Variable | Type | Description | Access Rate |
|----------|------|-------------|-------------|
| `credit_access` | categorical | Credit source | 27% total |
| `crop_insurance` | categorical | Insurance type | 6% covered |
| `climate_finance_access` | binary | Climate adaptation finance | 7% |

**Reflects gap**: Very low financial service access in rural SSA

### Inherited Variables (41 from Datasets 1-3)

**From Dataset 1 - Smallholder Farming (12 vars)**:
- Farm characteristics: AEZ, region, farm size, soil quality
- Climate: Annual & seasonal rainfall
- Demographics: Household size
- Market access & inputs: Distance, livestock, extension, fertilizer
- Yields: Maize production

**From Dataset 2 - Livestock (15 vars)**:
- Herd structure: Cattle, small ruminants, poultry
- Health: Disease incidence, vaccination, mortality
- Veterinary access: Distance, visits, treatment
- Management: Grazing, pasture, feed, water

**From Dataset 3 - Post-Harvest (14 vars)**:
- Storage: Type, duration, losses, hermetic tech
- Handling: Drying, pest control, sorting
- Markets: Sales, buyers, prices, transport

This creates a **comprehensive household profile** spanning production, livestock, post-harvest, and adaptation.

## Uses

### Direct Use

✅ **Permitted**:
- Climate adaptation research & algorithm development
- Technology adoption modeling
- Policy scenario analysis (with caveats)
- Agricultural extension program design
- Climate finance targeting research
- Educational training on climate-smart agriculture
- Resilience indicator development

❌ **Prohibited**:
- Real-world policy targeting of actual households
- Climate finance allocation decisions
- Technology subsidy targeting
- Any use that could harm vulnerable populations

### Research Applications

**Climate Adaptation Studies**:
- Modeling household adaptation responses to climate shocks
- Technology diffusion under resource constraints
- Migration as adaptation strategy
- Diversification patterns

**Agricultural Development**:
- Extension service optimization
- Technology adoption barriers
- Credit access impact on adaptation
- Digital agriculture potential

**Policy Analysis**:
- Climate finance targeting algorithms (test only)
- Agricultural insurance design
- Technology subsidy effectiveness
- Integrated adaptation pathways

## Dataset Creation

### Curation Rationale

Climate change is the defining challenge for African agriculture, yet **household-level adaptation data is scarce**:
- Privacy-sensitive (household finances, migration)
- Expensive to collect (multi-year, multi-topic surveys)
- Limited geographic coverage
- Difficult to share

This synthetic dataset enables **climate adaptation research without privacy concerns**.

### Source Data

#### Literature Sources (16 peer-reviewed papers for new variables)

**Climate Shocks**:
- Sheffield et al. (2014) - Drought monitoring SSA
- Paeth et al. (2011) - Flood exposure patterns
- Mbow et al. (2019) - IPCC Africa temperature trends
- Tall et al. (2018) - Climate information access

**Technology Adoption**:
- Suri (2011) - Improved seed adoption economics
- You et al. (2011) - Irrigation potential Africa
- Giller et al. (2009) - Conservation agriculture SSA
- Garrity et al. (2010) - Agroforestry adoption
- Diao et al. (2014) - Mechanization levels
- Tsan et al. (2019) - Digital agriculture uptake

**Adaptation Strategies**:
- Di Falco & Chavas (2009) - Crop diversification
- Barrett et al. (2001) - Livelihood diversification
- Gray & Mueller (2012) - Climate migration

**Support Access**:
- Fletschner & Kenney (2014) - Credit access rural women
- Carter et al. (2017) - Index insurance developing countries
- Buchner et al. (2019) - Climate finance landscape

**Plus 24+ sources** for inherited variables from Datasets 1-3.

### Personal and Sensitive Information

**None** - This dataset contains NO personal information. All data is synthetically generated.

## Bias, Risks, and Limitations

### Known Limitations

1. **Simplified climate-tech relationships**: Real adoption decisions more complex
2. **Cross-sectional only**: No panel dynamics (see Dataset 5 when available)
3. **Missing variables**: Social networks, labor availability, detailed prices
4. **No spatial coordinates**: Zone-level only
5. **Adoption barriers simplified**: Cultural, information, structural barriers not fully modeled

### Biases

- **Literature bias**: Reflects published research, may under-represent marginalized groups
- **Tech optimism**: May overstate technology effectiveness vs. literature
- **Temporal**: Primarily 2010-2020 patterns, climate change accelerating

### Recommendations

Users should:
1. **Always validate** on real climate adaptation data before deployment
2. **Acknowledge synthetic nature** in all publications
3. **Test sensitivity** to key assumptions (e.g., adoption rates)
4. **Consider context**: SSA-specific, may not transfer to other regions
5. **Cite properly** (see below)

## Additional Information

### Dataset Curators

**Electric Sheep Africa**
Generated using **Synthetic Data Generation Playbook** methodology

### Portfolio Context

This is **Dataset 4** in the African Agriculture Synthetic Data Portfolio:
1. ✅ [Smallholder Farming Systems](https://huggingface.co/datasets/electricsheepafrica/african-agriculture-synthetic)
2. ✅ [Livestock Health & Disease](https://huggingface.co/datasets/electricsheepafrica/livestock-health-disease-ssa-synthetic)
3. ✅ [Post-Harvest Value Chains](https://huggingface.co/datasets/electricsheepafrica/postharvest-value-chains-ssa-synthetic)
4. 🔄 **Climate Adaptation & Technology** (this dataset)
5. 📋 Panel Data (planned)

**Combined**: 4,000,000 records, 93 unique variables, 40+ literature sources

### Licensing Information

- **Dataset**: Creative Commons Attribution 4.0 (CC-BY-4.0)
- **Code**: MIT License
- **Documentation**: CC-BY-4.0

**Requirements**:
- ✅ Attribution required
- ✅ Must acknowledge synthetic nature
- ✅ Follow acceptable use policy
- ❌ No warranty provided

### Citation Information

```bibtex
@dataset{african_climate_adaptation_synthetic_2025,
title = {Climate Adaptation and Technology - Sub-Saharan Africa Synthetic Dataset},
author = {Electric Sheep Africa},
year = {2025},
publisher = {HuggingFace},
version = {1.0.0},
url = {https://huggingface.co/datasets/electricsheepafrica/climate-adaptation-tech-ssa-synthetic},
note = {Dataset 4 of African Agriculture Portfolio. 1 million synthetic records, 57 variables, 16 new literature sources for climate/tech variables}
}
```

**Cite the full portfolio**:
```bibtex
@collection{african_agriculture_portfolio_2025,
title = {African Agriculture Synthetic Data Portfolio},
author = {Electric Sheep Africa},
year = {2025},
publisher = {HuggingFace},
note = {4 datasets, 4 million records, 93 variables, 40+ literature sources}
}
```

### Contributions

Report issues or improvements:
- Data quality issues
- Documentation improvements
- New variable suggestions
- Validation findings against real climate adaptation data

## Quick Start

```python
# Load dataset
from datasets import load_dataset
ds = load_dataset("electricsheepafrica/climate-adaptation-tech-ssa-synthetic")

# Or with pandas
import pandas as pd
df = pd.read_parquet("hf://datasets/electricsheepafrica/climate-adaptation-tech-ssa-synthetic/climate_adaptation_data.parquet")

# Explore climate adaptation patterns
print(df[['drought_events_5yr', 'improved_seed_use', 'irrigation_access']].describe())

# Technology adoption by climate exposure
adoption_by_drought = df.groupby('drought_events_5yr')['improved_seed_use'].value_counts(normalize=True)
print(adoption_by_drought)

# Credit access and technology adoption relationship
credit_tech = pd.crosstab(df['credit_access'], df['conservation_agriculture'], normalize='index')
print(credit_tech)
```

## Validation

The dataset has been validated against:
- ✅ Literature benchmarks (adoption rates match published studies)
- ✅ Logical consistency (e.g., climate info access higher with extension)
- ✅ Conditional dependencies (e.g., irrigation higher in peri-urban)
- ✅ Missing data patterns realistic

See metadata files for detailed validation results.

---

**Questions?** See full documentation in repository
**Ready to use!** Load and start analyzing climate adaptation patterns immediately.

**Part of a larger portfolio**: Combine with Datasets 1-3 for integrated farm-livestock-postharvest-climate analysis.