Logo Lanfrica

EiA2030-ex-ante/Tanzania-Price-Predictions

Domaine:

agriculturesocioeconomic

Type de record:

dataset
Créateur:
EiA
Hôte:
This repository contains data and codes related to modelling and predicting food commodity prices for Tanzania. # Predictive mapping of wholesale crop prices for rural areas in Tanzania This repository contains the data, data preparation steps, and code used for a study on spatio-temporal prediction of wholesale crop prices for rural areas in Tanzania. This project builds and evaluates Random Forest models to predict monthly crop prices at high spatial resolution. The whole workflow used is compiled in this document- eia2030-ex-ante.github.io # Background Timely and location-specific data on wholesale crop prices in rural areas is often scarce in Tanzania. Available data is usually aggregated at the regional or national level (e.g., Dar es Salaam or zone-level averages), which fails to reflect the substantial price variability observed across rural markets. This lack of granular market information limits the ability of policymakers to respond effectively to local food price fluctuations, plan interventions, or invest in rural value chains. To address this, the study develops and tests a spatially explicit machine learning framework to predict monthly prices for eight key staple crops across rural Tanzania. These crops include maize, rice, sorghum, bulrush millet, finger millet, wheat, beans, and potatoes. The models use Random Forest algorithms trained on observed prices from 44 markets, leveraging environmental, temporal, and market-access covariates. Key predictors include proximity to cities and ports, population density, travel time, and bioclimatic variables. # Repository Structure # 1. Data This folder contains the cleaned market price data used for model training and validation. The raw data was sourced from the Tanzania Ministry of Industry and Trade (MIT), and includes monthly wholesale prices for key staple crops reported at various markets across the country. The cleaned dataset as is provided as a csv file 'Tanzania_Price_Data_AllCrops_with_Coordinates4.csv'. It covers 44 markets and includes these crops: maize, rice, sorghu …