Knowing the phytosanitary status of potato diseases in colombia using data science and multi-approach models

Abstract Text: Potato is a key crop in Colombia, contributing to food security and the rural economy. However, diseases threaten yields despite the expansion of production areas. In Colombia, although a protocol exists for phytosanitary data collection, its analysis is weak, leading to excessive fungicide use. Data science and modeling offer efficient ways to detect disease patterns, enabling evidence-based management. This study aimed to identify major potato diseases in Colombia and factors influencing their spread. The phytosanitary records of Federación Colombiana de Productores de Papa were used, which have information on the incidence, severity, management and technological status of productive lots between 2022 and 2024. Additionally, climatic information was acquired from databases such as NasaPower. Once the data had been processed and the most relevant diseases in the productive zones of the country had been identified, statical frequentist and Bayesian models, as well as machine learning were used to determine the distribution of the diseases, as well as the environmental and management factors that can influence their intensity measures. The results obtained allowed determining a first approximation of the phytosanitary status of the most relevant diseases in Colombia, and their variability at spatial, temporal and spatio-temporal levels. This information serves as a basis for the development of evidence-based prevention, control and management measures to improve crop yields in the country and reduce costs associated with inefficient management.