Peer review

Diverse on-farm agronomic data, gathered via precision agriculture (PA), require depuration and joint analysis of multiple georeferenced field characteristics such as crop yields, and soil and topographic aspects. Multivariate analysis of spatial variability has been recommended to understand variation of data within fields and classify sites into zones of broad similarity to support crop management. However, most of the statistical techniques available for spatial data require advanced skills. Cleaning and statistical analysis of big spatial data can only be effectively implemented if there are easy-to-use computer programs integrating the analytical steps. This paper explains the development and implementation of FastMapping, an interactive web application to automatically clean PA raw data, generate spatial variability field maps, and delineate multivariate management zones. The application uses an interface developed in R language that automatically depurates datafiles; in addition, through automatic spatial interpolation of each variable, it allows us to merge data layers on the same spatial grid in such a way that each site within the agricultural field has values of all the measured variables. On this grid, FastMapping identifies homogeneous zones, in a multivariate and spatial way, and provides data reports including validation of the delineated zones. FastMapping outputs for zone delineation were compared with results from Management Zone Analyst (MZA) software in several fields. The new software yielded a similar zoning to that of MZA and added graphs and statistical results that are lacking in most PA software tools. The flexibility of FastMapping to import, clean and visualize PA data in a single computer environment is highly encouraging. FastMapping can be used by the PA community to support site-specific agricultural management.

Text in English