Recipient Organization
UNIVERSITY OF CALIFORNIA, DAVIS
410 MRAK HALL
DAVIS,CA 95616-8671
Performing Department
Population Health & Reproduction
Non Technical Summary
Dairy production systems generates huge amounts of daily data at the cow level (e.g., milk production, components and quality, reproductive events, disease events, genetic information, among others), at the pen level (feed components and nutritional value, feed composition, amount fed), and farm level (revenues, cost of production, labor, maintenance, etc.). However, this data is not integrated, and farmers, managers, and consultants, miss opportunities for better decisions. In addition, in the era of big data and artificial intelligence, dairy researchers also miss opportunities to generate knowledge and information to improve the sustainability of the dairy industry. Because the dairy industry is a large part of the California economy, the ability to perform efficiency analysis at the farm level will enable the dairy industry to improve profitability, sustainability using economic and dairy model simulations will improve the California economy. This collaboration will also advance the knowledge on dairy data handling and integration and its applicability, which is currently a challenge in the dairy industry and provide training opportunities for graduate and veterinary students. Both UC Davis, students, and ultimately dairy industries, will benefit from this collaboration.
Animal Health Component
30%
Research Effort Categories
Basic
30%
Applied
30%
Developmental
40%
Goals / Objectives
The overall goal is to integrate large dairy data sets, create decision-making tools to be used in efficiency analyses and provide an environment in which researchers and graduate students in dairy and computer sciences can work with big data, computer modeling and efficiency analyses.OBJECTIVESThe objectives of this proposal are to 1. Modify an existing computer model of a dairy cow, Molly, to simulate pen level milk production and nutrient balance, and 2. Use this model and herd management data to perform within-farm efficiency analyses.
Project Methods
Design: 1. Develop algorithms to improve data integrity and functionality. Using a data from approximately 5 local dairies, procedures will be developed to gather, merge, and clean data from the herd management software (VAS). The value of databases is that data can be stored, manipulated, and queried to match. There is no known report on the quality of data that is gathered from dairy herds, and how the process of data collection can be improved to guarantee data integration and analysis.2. Modify Molly model to replicate the structure of a dairy, for instance the distribution of milk production within a pen and across pens and develop equations for efficiency analysis. Individual farm data will be integrated with Molly to replicate each sample dairy. Then the dairy will be measured against the model dairy from Molly (the optimal dairy) to help identify where efficiency could be improved. Such integration would allow farmers to identify opportunities for improving the efficiency of not only their nutritional, but also their environmental program by running efficiency analysis of their own herd against the outputs from Molly. In addition, there is no efficiency analysis algorithm built-in in herd management software. 3. Provide research and training for graduate and veterinary students on best practices for data collection, management and simulation and economic modeling techniques. These steps will allow further collaborations in research, training and extension projects to improve the scope of data integration, including dairy precision technologies, and the use of other tools such as artificial intelligence for on-farm management decisions.Data analysis: In process 1. (above), algorithms based on previous research projects will be developed further using Access databases and structured query language to process dairy data. In process 2., ability of the molly model to replicate the dairy will be assessed using root mean square prediction error, concordance correlation coefficient and the Nash-Sutcliffe efficiency method and their decompositions (Bibby and Toutenburg, 1977; Lin, 1989). Efficiency analysis will be performed using Data Envelope Analysis (Ferrel, 1957; Stokes et al., 2007) to identify less efficient pens, predict why they are less efficient and how improvement could increase profitability. Process 3. does not require data analysis.