BeefTrader Grass: an intelligence platform on market information to maximize profit of stock cattle

• Grant number: 19/09134-6
• Support type: Research Grants - Innovative Research in Small Business - PIPE
• Duration: March 01, 2021 - November 30, 2021
• Field of knowledge: Agronomical Sciences - Animal Husbandry - Animal Production
• Principal researcher: Murilo Garrett Moura Ferreira dos Santos
• Grantee: Murilo Garrett Moura Ferreira dos Santos
• Company: Tech - Inovações Tecnológicas para a Agropecuária Ltda
• CNAE: Criação de bovinos ; Atividades de apoio à pecuária ; Desenvolvimento de programas de computador sob encomenda
• City: Piracicaba
• Assoc. researchers: Alan Caio Rodrigues Marques ; Washington Alves de Oliveira
• Associated scholarship(s): 21/03600-5 - BeefTrader Grass: an intelligence platform on market information to maximize profit of stock cattle, BP.TT; 21/02924-1 - BeefTrader Grass: an intelligence platform on market information to maximize profit of stock cattle, BP.TT; 21/01572-4 - BeefTrader Grass: an intelligence platform on market information to maximize profit of stock cattle, BP.PIPE

Abstract

More than 90% of the Brazilian livestock is raised on pasture, making it the world's largest commercial herd that directly and indirectly moves more than R$ 597 billion (US$ 151,5 billion) or 8.7% of the Brazilian GDP. Extensive production systems have been losing their areas due to low profitability and agricultural expansion. This project aims to change this scenario completely, increasing and stimulating the agricultural activity by improving profitability and professionalism of its operations. For that purpose, this the algorithm will be developed (BeefTrader Grass) to maximize profit of stock cattle and finishing phases through early identification of animals with better performance. This initiative is a natural evolution of the award-winning, well-researched and successful BeefTrader platform supported by FAPESP (process: 2015/07855-7), honoring the efforts and resources deposited by the Foundation and Startup. All platform users that confine cattle, buy animals raised on pasture. These and other agents in the chain (beef and dairy farms, nutrition companies, animal health, animal breeding programs, and slaughterhouses) have consistently demanded a grass-fed intelligence platform from @Tech, enhancing the development of the current project. The research project will be divided into two stages: I) development of a mathematical programming algorithm, which will allow simulations of purchase of animals monitored and recreated on pasture for later selection of the most suitable ones to be finished. The objective of this algorithm is to maximize profitability by selecting more suitable animals from the stock cattle. A historical database (DB) of the stock cattle of partner farms (approximately 5,000 animals) weighed at an interval of 30-60 days will be used to implement the mathematical model. After the exploratory analysis (measurements of data position and dispersion) and study of atypical data, the DB will be organized. The hypothesis at this phase will be that BeefTrader Grass allows to increase the average profitability of selected animals when compared to the average profitability of the total population. II) Evaluation of the mathematical programming algorithm with an independent DB obtained from a field experiment. In this phase, the algorithm will be evaluated at least on three partner farms in three regions: Center-West, Southeast and South. At least 1,000 animals will be monitored as a group of contemporaries during the rearing stage. After collecting farm information (including economic and food management data), these animals will be identified and evaluated at the entrance to the farm (weight, race, gender, age, body structure degree, biometric measurements and Body Condition Score, BCS]. Throughout the breeding season, the animals will be monitored by daily body weighing. For each weighing event, the animals will be imaged by intelligent cameras (@Tech's 3DBeef system) thereby monitoring the BCS's rate of change automatically over time. During the stock cattle stage, with daily data from partner farms, it will be possible to test the hypothesis specified in Step 1 using BeefTrader Grass. Using lean development and user-based modeling, via interview, each farm will be visited during the experiment and on the departure date of the animals to evaluate and improve the platform. Expected products: 1) animal pasture performance variables (up to 60 d valuation); 2) research operations algorithms to maximize profitability between pasture stocker phase and feedlot; and 3) database from partners' pilot farms interviewed; 4) international BeefTrader Grass scientific presentation. (AU)