Source: PURDUE UNIVERSITY submitted to NRP
DATA SCIENCE AND ANALYTICS FOR PRECISION LIVESTOCK SYSTEMS
Sponsoring Institution
National Institute of Food and Agriculture
Project Status
ACTIVE
Funding Source
Reporting Frequency
Annual
Accession No.
1026642
Grant No.
2021-38420-34943
Cumulative Award Amt.
$241,000.00
Proposal No.
2021-03639
Multistate No.
(N/A)
Project Start Date
Jul 15, 2021
Project End Date
Jul 14, 2026
Grant Year
2021
Program Code
[KK]- National Needs Graduate Fellowships Program
Recipient Organization
PURDUE UNIVERSITY
(N/A)
WEST LAFAYETTE,IN 47907
Performing Department
Animal Science
Non Technical Summary
Rapid advancement of the Internet of Things and development of affordable livestock precision technologies are shaping a new era in food animal production. These tools generate complex datasets for precision management, optimization of breeding schemes, animal welfare monitoring, and comprehensive economic analysis. Big data analytics requires professionals with exemplary technical skills to make timely and accurate decisions that maximize the sustainability of livestock industries. Despite the widespread adoption of precision technology, currently generated datasets are being underutilized due to a deficit of qualified personnel with intersecting knowledge of data science and animal sciences. Purdue University is a leader in innovative teaching, research, and engagement in data science and digital agriculture. This is due to large investments in state-of-the-art facilities, close connections with industry partners, tight alignment with USDA strategic goals, and partnerships between multidisciplinary faculty. Therefore, the main objectives of this project are: 1) develop a Certificate in Livestock Data Science; and 2) recruit highly promising students (2 MS and 2 PhD) and train them in close partnership with private companies and livestock industry stakeholders. To promote diversity and inclusion, at least 75% of the students will be recruited from underrepresented minority groups or women. A multidisciplinary team will mentor these students and foster activities that provide strong training in both animal and data sciences as well as leadership, teaching, and mentoring skills. This project will increase the number and enhance diversity of livestock data scientists with the necessary technical and leadership skills to excel in either academia or industry.
Animal Health Component
(N/A)
Research Effort Categories
Basic
(N/A)
Applied
(N/A)
Developmental
(N/A)
Classification

Knowledge Area (KA)Subject of Investigation (SOI)Field of Science (FOS)Percent
2013410108010%
3033999108110%
3153999106010%
3013999106010%
0013999209010%
1323999208010%
1363999108010%
3023999101010%
9033999108110%
9033999208010%
Goals / Objectives
Rapid advancement of the Internet of Things and development of affordable livestock precision technologies are shaping a new era in food animal production. These tools generate complex datasets for precision management, optimization of breeding schemes, animal welfare monitoring, and comprehensive economic analysis. Big data analytics requires professionals with exemplary technical skills to make timely and accurate decisions that maximize the sustainability of livestock industries. Despite the widespread adoption of precision technology, currently generated datasets are being underutilized due to a deficit of qualified personnel with intersecting knowledge of data science and animal sciences. Purdue University is a leader in innovative teaching, research, and engagement in data science and digital agriculture. This is due to large investments in state-of-the-art facilities, close connections with industry partners, tight alignment with USDA strategic goals, and partnerships between multidisciplinary faculty. Therefore, the main objectives of this project are: 1) develop a Certificate in Livestock Data Science; and 2) recruit highly promising students (2 MS and 2 PhD) and train them in close partnership with private companies and livestock industry stakeholders. To promote diversity and inclusion, at least 75% of the students will be recruited from underrepresented minority groups or women. A multidisciplinary team will mentor these students and foster activities that provide strong training in both animal and data sciences as well as leadership, teaching, and mentoring skills. This project will increase the number and enhance diversity of livestock data scientists with the necessary technical and leadership skills to excel in either academia or industry.
Project Methods
SELECTED METHODS (PLEASE SEE FULL PROPOSAL FOR MORE DETAILS)Development of a Graduate Certificate in Livestock Data Science:Another important goal of this project will be the development of a Graduate Certificate in Livestock Data Science, which will expand based on an existing initiative. In order to obtain theGraduate Certificate in Livestock Data Science, the NNF-USDA fellows will be required to take a minimum of 20 credits and 5 credits from courses offered by the Dep. of Animal Sciences. Furthermore, they will be required to perform a research project based on livestock datasets, with direct applications to the animal science industry. An advisory committee formed by the co-PDs will meet twice in the first year to define all the coursework for each fellow. The same curricula requirements for obtaining the Graduate Certificate will be applied to both MSc and PhD fellows.Organization and philosophy of the proposed graduate training program:The fellows in this program will attain the core competencies for applying data science to the advancement of precision livestock production. This requires a multi-scale and cross-disciplinary collaboration among educators from different departments, including Animal Science, Computer Science, Agricultural and Biological Engineering, and Statistics. We will train students both in scientific competence in the aspects of livestock data science, and in the areas of leadership, communication, mentoring, and teaching. The team of co-PDs bring expertise in areas such as data acquisition, data wrangling, visualization, statistics and management decision making; their participation as co-advisors and advisory committee members assures a fuller experience for the fellows.Fellows will be expected to attain four core competencies: Specifically, fellows will have the capacity to: (1) conduct multidisciplinary research in livestock data science and communicate research results through peer-reviewed manuscripts, webinars, and presentations; (2) integrate electronic data collection from precision livestock farms, data analysis, and application skills from coursework; (3) employ leadership and management skills through participation in leadership and experiential-learning programs; and (4) effectively communicate their skills and technologies through informal and formal teaching through teaching assistantships, organization of workshops and Teaching Units, and invited guest lectures. The two PhD students will also participate in a Teaching Certificate Program from the Purdue Center for Instructional Excellence.Students will be expected to have a working knowledge of the entire data pipeline and be able to apply it to guide targeted actions to improve livestock efficiency, productivity, safety, and sustainability. Students pursuing a MS will be assessed through coursework, meetings with an academic committee (co-PDs), research performance evaluations, publication of manuscripts in high-impact factor journals, a public seminar, an overall assessment based on learning outcomes of the COA, and final defense. Assessment of PhD students will be the same as MSc students with the exception of a preliminary exam prior to candidacy for PhD students.Assessing, monitoring, counseling, guiding and sustaining Fellows' progress:Guidance will be provided through an academic advisory committee as well as peer-mentoring for the students. The academic advisors will initiate a meeting with the student to discuss expectations, set goals, and identify milestones and timeline to achieve these goals. Students will be required to complete training on theResponsible Conduct of Researchthrough CITI courses available at Purdue University. As part of this proposal, we will develop amultidisciplinary graduate student peer-mentoring program. Evidence suggests that peer mentoring is a valuable part of sustaining students through the rigors of a postgraduate training program (Lorenzetti et al., 2019). They will also have the opportunity to work collaboratively with the co-PDs of this project onco-mentoring undergraduate studentsin capstone research projects related to Livestock Data Science. In addition, we will initiate astudent-led monthly journal clubthat will expose students to a wide range of relevant research. Although student-led, this journal club will be guided by the co-PDs. Fellows will be required tojoin the active graduate student organization(University level) and will have the opportunity to be mentored through the Purdue Alliance for Graduate Education and the Professoriate (AGEP). Fellows will also be connected to appropriate cultural centers on campus.Professional mentoring:Fellows will be prepared to make immediate impact in careers in academia, industry or government (e.g., USDA-ARS; livestock breeding companies), and entrepreneurpositions in livestock data science and will receive professional mentoring.Experiential learning: hands-on research, extension, teaching and other activities:The educational activities related to this program will focus on preparing the next generation of graduates who will have anappreciation for all aspects of the land-grant mission. It is paramount that the students are well trained to integrate data science and analytics to advance livestock production in the US, especially those industries that are widely using precision livestock technologies. We will utilize annual farm and industry tours (with specific learning objectives defined by the co-PDs) to expose students to data needs in agriculture and as a way to increase exposure of students to potential careers in livestock data science. During these tours (> 4/year), students will be presented with a case study where they will discuss data needs to answer a problem in animal agriculture. Students will reflect on their experiences by presenting information to Data Mine students in our Outside Events program, and/or when they are guest lecturing in animal production classes. All students will be required to organize and teach a hands-on, student-centered 12-hour workshop or Teaching Unit for undergraduate students (supervised by the co-PDs) related to the use of big data in animal agriculture.Our overall goal is to utilize a hands-on, experiential approach, to create greater interest and engagement in data science in animal agriculture. By increasing interest and exposure through multiple experiences, students will be more likely to remain in the field after graduation.Data manipulation, visualization, and informed decision making will be key priority of the training of these students. Each of the students involved in this project will have access to large-scale, high-throughput phenotypes from precision livestock farms and private tech companies. In addition to their own research, the students will work on practical case studies of livestock data mining. All the results obtained will be presented and discussed with the stakeholders to foster translation of results, but to also integrate end-user input to ensure relevance. Such interactions will also offer students opportunities to start interacting with their potential employers, learn their needs, and the skills they find most valuable in new employees. Based on the extensive track record of related projects, a key goal in this project is to enable the students to be fully integrated in Livestock Data Science research endeavors with their research mentor, including research dissemination in journal papers, conference publications, and posters. Students recruited through this project will interact routinely with students from the Data Mine, OATS Center, and other data science and agricultural disciplines. This will be a great opportunity for: i)knowledge exchange, ii) helping students to build mentorship/leadership skills, and iii) recruiting new undergrads with interest in Livestock Data Science.

Progress 07/15/23 to 07/14/24

Outputs
Target Audience:During this reporting period, our main target audience was the NNF fellows (graduate students) hired at Purdue University and other graduate students from Purdue University. They were trained and interacted with faculty across various disciplines, including animal sciences, computer science, statistics, veterinary science, and entomology. The activities performed by them also reached various audiences such as >300 undergraduate students in the Purdue Department of Animal Sciences, data science organizations in which they did their internships at, and other data science stakeholders that the faculty and the NNF fellows interacted with such as private companies, Purdue Ag Data Services, Lely, Afimilk, Zoetis, and DairyComp. In addition, the first publications from the NNF fellows have been presented at national and international conferences and also through scientific publications. Therefore, our project has had a broader impact in the scientific communities in animal and data sciences. Changes/Problems: Nothing Reported What opportunities for training and professional development has the project provided?The NNF fellows have attended conferences in the area of Livestock Data Science in Tennessee and in Calgary (Canada), completed multidisciplinaryclasses across departments, presented seminars, participated in internships in the data science organizations, performed multidisciplinary research projects, and presented and published the results of their research. Two fellows have also completed their programs and were employed in the area of livestock data science, which is a fantastic outcome of this project. How have the results been disseminated to communities of interest?The results have been disseminated through conference and journal publications, oral presentation, and graduate certificate developed through this program. What do you plan to do during the next reporting period to accomplish the goals?All the remaining graduate students will continue their training, research projects, internships, and other planned activities. The Graduate Certificate in Livestock Data Science has been launched and more students will start enrolling on it in the coming semesters.

Impacts
What was accomplished under these goals? We have recruited and trained 4 (all female) graduate students who started their graduate programs at Purdue University.Dr. Jackie Boerman, Dr. Allan Schinckel, and Dr. Luiz Brito are serving as their major advisors and other Co-PIs will be part of their advisory committee. We have also designed a Graduate Certificate in Livestock Data Science, which was launched this semester. A brief description of the Graduate Certificate is provided below. The fellows have also completed all the activities planned such as presenting seminars in the area of livestock data science, are conducting their research projects, have completed internships in data science organizations, and met all the curse work requirements from their graduate programs. They have also published various results of their projects in national and international conferences and in per-reviewed journals. Graduate Certificate in livestock Data Science:This Graduate Certificate in Livestock Data Science will provide various complementary opportunities for Purdue graduate students (at both MS and PhD levels) to improve their knowledge and data science skills in the context of sustainable livestock production. This certificate will consist of a minimum of 12 graded credits (50000 or 60000 level courses) chosen from three main course groups: ≥ 3 credits from Group 1 (Animal Sciences), ≥ 6 credits from Group 2 (Computing and Data Science), and ≥ 3 credits from Group 3 (Applied Statistics). A detailed list of the courses is presented as Appendix and will be updated as courses are created or removed from the Purdue University catalogue. Students will also be required to successfully complete two semesters of a 1-credit Livestock Data Science Seminar (Pass/Fail evaluation; in addition to the 12 credits of letter graded courses). Furthermore, the students will be required to perform at least one research project (equivalent to a thesis chapter or scientific paper) using large-scale livestock datasets, with direct applications to the animal industry. During their program, all students will be required to organize and teach a hands-on >4-hour workshop (supervised by Purdue faculty and coordinated by Dr. Elizabeth Karcher) related to the use of big data in animal agriculture. The audience of these workshops will be undergraduate students or other agricultural stakeholders (extension workshops). Lastly, all the students pursuing the Graduate Certificate in Livestock Data Science will be required to complete at least 80 hours of internship in a data science organization at Purdue University (e.g., Agriculture Data Services, The Data Mine) or in a livestock company (in consultation with their advisory committee and Graduate Certificate Program Coordinator).

Publications

  • Type: Journal Articles Status: Published Year Published: 2024 Citation: Wonders O. Ogundare, Linda M. Beckett, Leriana G. Reis, McKeeley C. Stansberry, Sydney N. Roberts, Uchenna Y. Anele, Allan P. Schinckel, Theresa M. Casey, Radiah C. Minor. The impact of cooling and Moringa supplementation on oxidative stress in serum and milk, including milk cytokines, in heat stressed lactating sows and their litters. Translational Animal Science Journal. (Accepted)
  • Type: Journal Articles Status: Accepted Year Published: 2024 Citation: 2) McKeeley C. Stansberry, Wonders Ogundare, Linda M. Beckett, Leriana Garcia Reis, Evy M. Tobolski, Brian T. Richert, Theresa M. Casey, Allan P. Schinckel, Radiah C. Minor. The effect of electronic cooling pads and Moringa oleifera supplementation from late gestation to weaning on sow production performance under heat stress conditions. Journal of Animal Science. Accepted
  • Type: Journal Articles Status: Submitted Year Published: 2024 Citation: 3) McKeeley C. Stansberry, Wonders Ogundare, Linda M. Beckett, Leriana Garcia Reis, Elizabeth G. Fisher, Evy M. Tobolski, Brian T. Richert, Theresa M. Casey, Allan P. Schinkel, Radiah C. Minor. The effect of cooling and Moringa oleifera supplementation from late gestation to weaning on sow milk lipidome. Journal of Animal Science. Submitted.
  • Type: Conference Papers and Presentations Status: Published Year Published: 2024 Citation: Near-real-time feed bunk monitoring for dairy cattle using stereo vision. M. N. Flinders, P. Rao, D. R. Buckmaster, A. R. Reibman, and J. P. Boerman - Presented at ADSA 2024. Florida, USA
  • Type: Conference Papers and Presentations Status: Published Year Published: 2024 Citation: Stereo vision to predict dry matter intake, M. N. Flinders, P. Rao, A. R. Reibman, and J. P. Boerman  Presented at the Tri-State Dairy Nutrition Conference 2024
  • Type: Conference Papers and Presentations Status: Published Year Published: 2024 Citation: Rao, P., M.N. Flinders, D. Buckmaster, A.R. Reibman, and J.P. Boerman. 2024. Real-time Cattle Intake Monitoring Using Stereo Vision. Electronic Imaging 36:16.  Presented at Imaging Conference but it is also a paper (they make them write the paper right after the conference).
  • Type: Conference Papers and Presentations Status: Published Year Published: 2024 Citation: Moulder, C.N., Maskal, J.M., Kalbaugh, K., Pacheco, H., Boerman, J. and Brito, L.F., 2024. 87 Estimation of genetic parameters for activity levels in North American Holstein cattle. Journal of Animal Science, 102(Supplement_3), pp.23-24.
  • Type: Conference Papers and Presentations Status: Published Year Published: 2024 Citation: McKeeley C Stansberry, Wonders Ogundare, Linda M Beckett, Leriana Garcia Reis, Elizabeth G Fisher, Evy M Tobolski, Brian T Richert, Theresa M Casey, Allan P Schinckel, Radiah C Minor, 287 The effect of cooling pad and Moringa oleifera on lactating sows under heat stress, Journal of Animal Science, Volume 102, Issue Supplement_2, May 2024, Pages 228229, https://doi.org/10.1093/jas/skae102.259
  • Type: Conference Papers and Presentations Status: Published Year Published: 2024 Citation: Eimear C Mannion, Theresa Casey, Allan P Schinckel, Jung Yeol Sung, Wonders Oogundare, McKeeley Stansberry, Robert M Stwalley, Radiah Minor, Samantha Neeno, Tyler C Field, PSIV-6 Impact of Cooling Pads and Dietary Moringa on Measures of Heat Stress in Sows in Late Gestation, Journal of Animal Science, Volume 101, Issue Supplement_2, November 2023, Pages 348349, https://doi.org/10.1093/jas/skad341.396


Progress 07/15/22 to 07/14/23

Outputs
Target Audience:During this reporting period, our main target audience was the NNF fellows (graduate students) hired at Purdue University. They were trained and interacted with faculty across various disciplines, including animal sciences, computer science, statistics, veterinary science, and entomology. The activities performed by them also reached various audiences such as >400 undergraduate students in the Purdue Department of Animal Sciences, data science organizations in which they did their internships at, and other data science stakeholders that the faculty and the NNF fellows interacted with such as private companies, Purdue Ag Data Services, Lely, Afimilk, Zoetis, and DairyComp. Changes/Problems:All the activities are being developed as planned. What opportunities for training and professional development has the project provided?The NNF fellows have attended conferences in the area of Livestock Data Science in Tennessee and in Calgary (Canada), completed multidisciplinary classes across departments, presented seminars, participated in internships in the data science organizations, and performed multidisciplinary research projects. How have the results been disseminated to communities of interest? Nothing Reported What do you plan to do during the next reporting period to accomplish the goals?All the graduate students will continue their training, research projects, internships, and other planned activities. The Graduate Certificate in Livestock Data Science has been approved at two levels and is now at the last stage.

Impacts
What was accomplished under these goals? We have recruited all the 4 (all female) graduate students who started their graduate programs at Purdue University.Dr. Jackie Boerman, Dr. Allan Schinckel, and Dr. Luiz Brito are serving as their major advisors and other Co-PIs will be part of their advisory committee. We have also designed a Graduate Certificate in Livestock Data Science, which was approved by the Department of Animal Sciences and by the Purdue College of Agriculture. A brief description of the Graduate Certificate is provided below. The fellows have also completed all the activities planned such as presenting seminars in the area of livestock data science, are conducting their research projects, have completed internships in data science organizations, and met all the curse work requirements from their graduate programs. Graduate Certificate in livestock Data Science:This Graduate Certificate in Livestock Data Science will provide various complementary opportunities for Purdue graduate students (at both MS and PhD levels) to improve their knowledge and data science skills in the context of sustainable livestock production. This certificate will consist of a minimum of 12 graded credits (50000 or 60000 level courses) chosen from three main course groups: ≥ 3 credits from Group 1 (Animal Sciences), ≥ 6 credits from Group 2 (Computing and Data Science), and ≥ 3 credits from Group 3 (Applied Statistics). A detailed list of the courses is presented as Appendix and will be updated as courses are created or removed from the Purdue University catalogue. Students will also be required to successfully complete two semesters of a 1-credit Livestock Data Science Seminar (Pass/Fail evaluation; in addition to the 12 credits of letter graded courses). Furthermore, the students will be required to perform at least one research project (equivalent to a thesis chapter or scientific paper) using large-scale livestock datasets, with direct applications to the animal industry. During their program, all students will be required to organize and teach a hands-on >4-hour workshop (supervised by Purdue faculty and coordinated by Dr. Elizabeth Karcher) related to the use of big data in animal agriculture. The audience of these workshops will be undergraduate students or other agricultural stakeholders (extension workshops). Lastly, all the students pursuing the Graduate Certificate in Livestock Data Science will be required to complete at least 80 hours of internship in a data science organization at Purdue University (e.g., Agriculture Data Services, The Data Mine) or in a livestock company (in consultation with their advisory committee and Graduate Certificate Program Coordinator).

Publications


    Progress 07/15/21 to 07/14/22

    Outputs
    Target Audience:During this first year of the project, we reached out to potneital graduate students to join our National Needs Fellowship program at Purdue University. As planned in the project, we contacted numerous Minority-serving institutions and other minority academic organizations across the country. The communication was done via email, phone calls, advertisement in email listservs, and communication in conferences and seminars in which the investigators were present. Therefore, the main target of this first year of the project was prospective graduate students. All the other activities were done internally such as the development of a Graduate Certificate in livestock Data Science and a new graduate course. Changes/Problems:There were no major changes or problems in this first year of the project. We aimed to increase diversity in the field of Data Science and although we had no minority applicants, all the three fellows hired are women. We are still recruiting the last fellow and we have very strong prospective candidates. What opportunities for training and professional development has the project provided?The training opportunities will start in the Fall/2022 when our three NNF felllows start their programs. How have the results been disseminated to communities of interest? Nothing Reported What do you plan to do during the next reporting period to accomplish the goals?During the next phase of the project, we will have one additional PhD student, will complete the approval process of the Gradaute Certificate in Livestock Data Science, and will establish the training activities that the students will be involved in, including a Data Science Journal Club and a Data ScienceInternship Program.

    Impacts
    What was accomplished under these goals? We have recruited three female graduate students who will start their programs on August 15, 2022. Dr. Jackie Boerman, Dr. Allan Schinckel, and Dr. Luiz Brito will serve as their major advisors and other Co-PIs will be part of their advisory committee. We have also designed a Gradaute Certificate in Livestock Data Science, which has been aproved by the Department of Animal Sciences and is now under evaluation by the Purdue College of Agriculture. A brief description of the Gradaute Certificate is provided below. Graduate Certificate in livestock Data Science:This Graduate Certificate in Livestock Data Science will provide various complementary opportunities for Purdue graduate students (at both MS and PhD levels) to improve their knowledge and data science skills in the context of sustainable livestock production. This certificate will consist of a minimum of 12 graded credits (50000 or 60000 level courses) chosen from three main course groups: ≥ 3 credits from Group 1 (Animal Sciences), ≥ 6 credits from Group 2 (Computing and Data Science), and ≥ 3 credits from Group 3 (Applied Statistics). A detailed list of the courses is presented as Appendix and will be updated as courses are created or removed from the Purdue University catalogue. Students will also be required to successfully complete two semesters of a 1-credit Livestock Data Science Seminar (Pass/Fail evaluation; in addition to the 12 credits of letter graded courses). Furthermore, the students will be required to perform at least one research project (equivalent to a thesis chapter or scientific paper) using large-scale livestock datasets, with direct applications to the animal industry. During their program, all students will be required to organize and teach a hands-on >4-hour workshop (supervised by Purdue faculty and coordinated by Dr. Elizabeth Karcher) related to the use of big data in animal agriculture. The audience of these workshops will be undergraduate students or other agricultural stakeholders (extension workshops). Lastly, all the students pursuing the Graduate Certificate in Livestock Data Science will be required to complete at least 80 hours of internship in a data science organization at Purdue University (e.g., Agriculture Data Services, The Data Mine) or in a livestock company (in consultation with their advisory committee and Graduate Certificate Program Coordinator). ?

    Publications