Smallholder Farmer Behaviour & Decision Dataset
20K+ longitudinal household survey records tracking smallholder farmer decision-making across crop selection, input adoption, market channel preference, and climate-adaptation strategy — the behavioural foundation for personalised agri-advisory and extension AI across Nigeria, Ghana, Tanzania, and Ethiopia.
This is a synthetic dataset generated from high-quality expert-labelled seed data. All records are algorithmically derived — statistical distributions, inter-field correlations, and annotation characteristics faithfully replicate real-world patterns from the source data, while ensuring no real individual, organisation, or transaction can be identified or reconstructed.
The African Smallholder Farmer Behaviour & Decision Dataset captures the decision-making patterns of 20K+ farm households across Nigeria, Ghana, Tanzania, and Ethiopia through repeated household surveys conducted over three consecutive growing seasons (2021–2024). Each record encodes the farmer's crop-portfolio choices, input usage decisions (fertiliser, improved seed, pesticide), market channel preference (farmgate, local market, cooperative, digital platform), and perceived climate-risk adaptation strategies.
Survey instruments were designed in close collaboration with national agricultural extension services and draw on established behavioural frameworks including the Technology Acceptance Model, Prospect Theory, and the Sustainable Livelihoods Framework. Enumerators were trained to administer the questionnaire in eight local languages; translated response values are harmonised to English-language enums in the dataset. Household wealth proxy scores are computed from asset indices validated against DHS-comparable survey rounds.
The dataset is structured for panel analysis — each household carries a persistent anonymised ID, enabling longitudinal modelling of adoption curves, yield-behaviour correlations, and the effect of weather shocks on decision-making. Advisory AI teams have used it to build farmer-segment classifiers, personalised input-recommendation engines, and extension-service targeting models that outperform blanket-broadcast advisory approaches by 30–40 % on adoption rate metrics in pilot evaluations.
Key Use Cases
Data Quality Indicators
Geographic Coverage
Dataset Schema
Each record represents one household survey response for a single growing season. Fields cover household identity, farm characteristics, crop decisions, input use, market behaviour, climate perception, and annotated segment labels.
| Field Name | Type | Description | Nullable | Example |
|---|---|---|---|---|
| household_id | STRING | Anonymised persistent household identifier | No | HH-NGA-TG-08814 |
| country_code | STRING | ISO 3166-1 alpha-2 country code | No | NG |
| admin1 | STRING | First-level administrative division | No | Taraba |
| survey_season | STRING | Survey growing season (YYYY-S1 or YYYY-S2) | No | 2023-S1 |
| farm_size_ha | FLOAT | Total cultivated farm area in hectares | Yes | 1.5 |
| primary_crop | ENUM | Main crop grown this season: MAIZE, CASSAVA, SORGHUM, RICE, YAM, OTHER | No | MAIZE |
| num_crops_grown | INTEGER | Total number of distinct crops grown this season | No | 3 |
| improved_seed_used | BOOLEAN | True if certified / improved seed variety was used | No | true |
| fertiliser_used | BOOLEAN | True if chemical or organic fertiliser was applied | No | false |
| pesticide_used | BOOLEAN | True if pesticide or herbicide was applied | Yes | false |
| market_channel | ENUM | Primary sales channel: FARMGATE, LOCAL_MARKET, COOPERATIVE, DIGITAL | No | LOCAL_MARKET |
| has_smartphone | BOOLEAN | True if the farmer owns or has regular access to a smartphone | No | true |
| uses_agri_app | BOOLEAN | True if the farmer uses any mobile agri-advisory application | Yes | false |
| extension_contact_frequency | ENUM | How often the farmer contacts extension services: NEVER, ANNUAL, SEASONAL, MONTHLY | No | SEASONAL |
| climate_risk_perception | ENUM | Perceived climate risk level: LOW, MODERATE, HIGH, VERY_HIGH | No | HIGH |
| adaptation_strategy | ENUM | Primary adaptation: NONE, CROP_SWITCH, IRRIGATION, EARLY_PLANTING, DIVERSIFICATION | Yes | EARLY_PLANTING |
| wealth_index_score | FLOAT | Composite household wealth index (0–100, DHS-validated) | Yes | 41.2 |
| household_head_gender | ENUM | Gender of household head: MALE, FEMALE | No | MALE |
| years_farming | INTEGER | Years the household head has been farming | No | 14 |
| farmer_segment | ENUM | Annotated segment: SUBSISTENCE, TRANSITIONAL, COMMERCIAL | No | TRANSITIONAL |
Sample Records
Four representative household survey records spanning countries, farmer segments, and climate adaptation strategies.
Build with Data that reflects Africa
Request access to our full catalog of licensed human-validated African datasets or request custom data tailored to your project.