Food Market Price & Supply Chain Dataset

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 Food Market Price & Supply Chain Dataset aggregates 15 years of price monitoring data from wholesale and retail markets in Nigeria, Ghana, Kenya, Ethiopia, and Senegal — five of the continent's most strategically important food economies. Each record captures the observed price of a specific commodity at a named market on a given date, alongside matched context variables: seasonal rainfall anomaly, national fuel price index, and a transport-cost proxy derived from road-condition and distance data.

Commodities covered include the core staple basket: maize, rice, sorghum, millet, cassava flour, yam, beans, soybean, wheat flour, and palm oil — 10 commodities × 5 countries × roughly 780 weekly observations per series, yielding a dense multi-market panel. Price data are sourced from national food-price monitoring systems (WFP VAM, FEWS NET, national statistics bureaux) and harmonised to USD-per-kg equivalents using contemporaneous exchange rates. Missing weeks are flagged with an imputation method code rather than interpolated silently.

The dataset is designed for time-series forecasting, causal inference (impact of fuel shocks on food prices), and spatial price-transmission modelling. It is structured as a long-format panel compatible with standard econometric and ML time-series libraries (statsmodels, Prophet, Darts, NeuralForecast). A companion GeoJSON file maps market locations for spatial interpolation tasks.