A dataset for soybean yield prediction (Machine Learning applications)
Description
The dataset combines meteorological, soil and soybean yield data collected from three publicly available sources: the Brazilian National Institute of Meteorology (Instituto Nacional de Meteorologia - INMET), the Brazilian Institute of Geography and Statistics (Instituto Brasileiro de Geografia e Estatística - IBGE), and the International Soil Reference and Information Centre (ISRIC). The study focuses on the Brazilian Center-West region, comprised the states of Mato Grosso do Sul (MS), Mato Grosso (MT) and Goiás (GO), since this region plays a relevant role in the national soybean production. The final dataset consists of 1056 observations and 37 variables. Each observation corresponds to a municipality–harvest instance and is associated with the soybean yield observed at the end of the growing season. Meteorological variables were derived from measurements collected by INMET every 15 minutes and aggregated using simple averages over the October–December period. This aggregation captures the climatic conditions experienced during a substantial portion of the soybean growing cycle (approximately 125 days) prior to harvest. A two-month forecasting horizon has been adopted in previous crop yield prediction studies. Soil variables were collected at the municipality level and used as static features, while additional engineered variables were extracted from the raw meteorological data. During the preparation of this work, the author(s) used Claude to organize, review and improve coding, after the rationale and the main instances, functions and logic has already been defined. After using this tool/service, the author(s) reviewed and edited the content as needed and take(s) full responsibility for the content of the published dataset.