Lokoja

Lokoja Intelligence (lokoja.app) — 'the price intelligence layer for Africa's food trade' — Two deliberately separated datasets. (1) PRICE/COST LAYER: end-to-end trade pricing across 67 Nigerian markets for 8 staples, built from three labelled layers — World Bank Real-Time Food Prices (model-estimated, monthly, back to 2007), WFP market monitoring (enumerator-observed, monthly rotating, plus a Kenya layer), and Lokoja's own weekly field ambassadors. Crucially it prices the trade, not just the commodity: corridor spreads quoted NET OF HAULAGE using a fixed loading cost plus per-tonne-km trucki…

Lokoja offers (Alternative, Price, Supply Chain, Geospatial, Sentiment) — Stated: 67 Nigerian markets x 8 staples with continuous coverage back to 2007 on the World Bank layer (~19 years), plus monthly rotating WFP observations, plus weekly ambassador reports from 3 markets, plus a transport-cost and corridor-spread layer over the published corridors. The long-run price panel is therefore substantial (tens of thousands of market-month observations), while the genuinely proprietary layers are thinner — weekly ambassador series from 3 markets to date, and a demand-signal log the operator concedes is only weeks-deep..

Physical commodity arbitrage and trade desk execution — transport-netted, realizable corridor spreads are the rare dataset that tells a trader whether a spread is actually bankable rather than nominally wide; intra-African food logistics cost modelling with a diesel-indexed per-tonne-km curve calibrated to real transporters rather than assumed rates; country risk pricing from the conflict-corridor exclusions; food-inflation nowcasting triangulating model-estimated against observed layers; supply-condition early warning from weekly ambassador supply levels, disruption reports and trader counts; and demand-intent intelligence from the WhatsApp stream — who is asking about what commodity, at what trade size, in which location and language, which is pre-purchase demand that transaction data cannot show.

The data is with 19 years of history.