The recommendations of liming and fertilization for the different crops in the main producing regions of the country are based on the knowledge generated by researchers in the area of soil science. Interpretation tables for nutrient contents in the soil and tables with the amount of nitrogen, phosphorus and potassium were elaborated for the different manuals published in some states / regions. Under the name of AGROFERT, a WEB system was created that aims to integrate the information of the correction and fertilizer recommendation systems for the states of Rio Grande do Sul / Santa Catarina and Paraná for the main grain crops and to evaluate the evolution of the characteristics of the soil. The system has a consistent scientific basis, allowing for more comprehensive recommendations that can be continually improved. To date, different existing computerized recommendation systems are specific to a single state / region and provide simpler information. The Agrofert system saves time, since its use does not require calculating, correcting and writing recommendations by the traditional method (Manual)

No abstract

The agribusiness has a significant importance to the Brazilian economy. In this case, technologies are needed to increase the efficiency of the productive system, but there are few technological tools that perform interpretations of soil analysis and fertilizer recommendations efficiently and comprehensively. In this context, the AgroFert system proposes to make corrective and fertilizer recommendations based on soil analyzes for several regions and grain crops of Brazil, through two systems: web and mobile, which uses a knowledge base in liming and fertilization. Therefore, this work aims at presenting the conceptual and logical modeling process of a database that have rules of information management to interpret the results of soil analysis reports and to recommend soil correctives and fertilizers. Rules were created that use comparisons, filtering, mathematical operations and logical operations, modeled in the SQL language, and when executed, generate interpretations of soil analysis and recommendations for correctives and fertilizers. The AgroFert covers the states of Rio Grande do Sul, Santa Catarina, Paraná, Minas Gerais and São Paulo and the biome of Cerrado. In addition to the conceptual modeling, 22 operations were created, totaling approximately 400 lines of code, and it was possible to recommend fertilizers for 13 grain crops. The results of the tests showed that the computerization of the data present in the RS / SC and Paraná manuals facilitated the interpretation of the data present in soil analysis reports, besides increasing the precision and agility in the recommendations of corrective and fertilizer producers.