Recipient Organization
UNIVERSITY OF FLORIDA
G022 MCCARTY HALL
GAINESVILLE,FL 32611
Performing Department
Agricultural and Biological Engineering
Non Technical Summary
Big data enriches the in-depth, game-change analytics to be at a scale and speed that the old modelling approach won't appropriate (Devlin, 2012). It has shown the potential for efficiency upgrade in many industries, such as banking, insurance, legal, and media. Yet, its application in agriculture is still at an early stage (Kamilaris, Kartakoullis, & Prenafeta-Boldu, 2017). Smart agriculture, as a cyber-physical system powered by IoT, remote sending, and cloud computing, will be a major opportunity for big data application (Friess, 2016). The potential benefit should not only be considered to increase the production but also improve the profitability of the entire supply chain and address the food security concerns (Esmeijer, Bakker, Ooms, Kotterink, & van Economische Zaken, 2015; Kamilaris et al., 2017).The motivations of applying big data analytics in managing agricultural or related systems come from different aspects. From economic perspective, business practitioners (e.g. growers, sellers) within the supply chain aim to grow the entire market by obtaining a better price while reducing cost at the same time. Since the advancement of the general knowledge derived from research experiments is a cumulative progress, extra local or field specific information could make a big margin under the assistance by big data technologies (Poppe, Wolfert, Verdouw, & Renwick, 2015; Sonka, 2015). Moreover, the general public is paying more attentions on global food security (Gilpin, 2014) and health related local food safety (Liu, Tan, & Zhang, 2015). High tech industry is another driver of big data analytics in agriculture. IoT and wireless data transfer enables remote sensing data collection (Kim, Lee, & Shin, 2018); cloud technology serves as a platform for high volume data storage and processing (Waga & Rabah, 2014); different end points (e.g. mobile apps, websites, robots) create interactions between users and backend systems (Bartlett, Andales, Arabi, & Bauder, 2015).Understanding the potential benefits and opportunities, the application of big data analytics in agriculture also encounter several barriers.Data ownership: As mentioned previously, data has become the extra resource that creates margin in the whole industry. Although necessary, a general agreement on fair data sharing across different stakeholders usually faces difficulties due to high cost in data collection (Kempenaar et al., 2016). "How to explore the full value of data through proper usage under a structure or governance" should be considered in any application (Nandyala & Kim, 2016).Data integration: Agricultural data are known to be highly heterogeneous (Xiu-feng, Shou-he, & Lei-feng, 2014) due to different data collection methods and sources. Integrating this information is the first prerequisite for a business model or system simulation, but always case specific.Data quality: Technical challenges, harsh environment, social uncertainties, etc., involved in remote sensing process create big challenges in the reliability and usability of data in agriculture. Common issue includes: missing data, inconsistent time step, low resolution, etc. (Campbell et al., 2013; Durre, Menne, Gleason, Houston, & Vose, 2010; Madsen, 1989; Napoly, Grassmann, Meier, & Fenner, 2018).Technology: Achieving smart agriculture is a technology intensive work. IT infrastructure that supports the collection, management and analysis of big data requires specific expertise (Frelat et al., 2016; Sawant, Urkude, & Jawale, 2016). In addition, appropriate applications of cutting-edge intelligence technologies, e.g. machine learning, artificial intelligence, predictive analysis, etc., usually rely on inter-discipline background and teamwork.This research aims to address these issues from policy and technology aspects. A policy framework that governing the data management and security is essential in protecting owners' copyrights and ownership from user access. Extension education is needed to help the farmers and data users build the knowledge and understanding of the value and restrictions in sharing and utilizing big data in agriculture. Compiling data from different formats should be pipelined by generating new standards or adopting existing ones, such as Open Geospatial Consortium (OGC). Advanced technologies and methodologies should be embedded into a sophisticated quality assurance and quality control (QAQC) procedure for both data providers and data users to mitigate the information loss and bias introducing during data collection and share. Cloud technology is critical to simplify the construction and maintenance of IT architecture for data management. Various analysis methodologies, traditional and cutting-edge, corresponding to different data types need to be fully explored, while the final modelling should be built on solid understanding of the physical mechanism and relationships of agriculture and related disciplines.A search of peer-reviewed publications will be conducted for applications or trails of data governance, integration standards, QAQC methods, cloud technologies and AI in different aspects of agriculture. The principal investigator has the data science experiences from both industrial and academia to lead and train the team and conduct multidisciplinary collaborative research in this area in addition to access to resources to accomplish the work. The impacts of the work are wide-reaching. Data providers, producers, modelers, end users, investors, agricultural planners, and policymakers can use the products to make informed decisions and choices.
Animal Health Component
30%
Research Effort Categories
Basic
40%
Applied
30%
Developmental
30%
Goals / Objectives
The overall goal of this research project is to investigate agriculture intelligence through data driven analytics to improve outcomes from agricultural producers and increase the efficiency of using natural and virtual resources. Specific objectives of this research include a focus to:Strengthen the data management life cycle of agriculture from data collection, transferring, storage, process and productization in different operation scalesGenerate standards or governance for ensuring the ownership and the fair sharing of the data collected from various sourcesGenerate executable and scalable data curation processes to improve the quality, reliability and usability of data for both data providers and usersDetect, analyze and interpret the physical mechanisms in agriculture by data from various sources and formatsDevelop smart management strategies for agriculture practices to improve the productivity and efficiency in resource utilizationAssess the impact of future challenges (e.g. climate change, population growth) on agriculture using innovative technologiesBuild databases or repositories to store, integrate and share the information for specific agriculture topics, such as benchmarks for AI in recognize tree fruit objectives, weather forecast in monitoring network data, soil moisture nutrient response to plant growth, etc.Build frameworks for extracting, storing and querying agriculture related knowledge through network-like structure for quick knowledge deliveryExplore cutting edge technologies in data science, statistics and mathematics to address the problems in agriculture data analyticsIntroduce the engagements from the general public to involve the decision-making, planning and maintenance of agriculture projects by developing interactive mediated decision support platforms for both rural and urban settings
Project Methods
The methods of this research project correspond to the stated overall goal and objectives. Although this project will initially focus on investigate agriculture knowledge, technology and applications in the state of Florida, the scope will likely expand geographically when further collaborations established with other entities, such as universities, government agents and private companies. Anticipated methods of this project:Collaborate with researchers, industry stakeholders, students, government agencies, and non-government agencies to identify issues and needs that can potentially be addressed with big dataConduct a systematic literature review of data governance, integration standards, QAQC methods, cloud technologies and AI to push the boundaries of current knowledge.Collaborate on and develop white papers to provide information or guide readers on issues related to agriculture data analyticsUse knowledge gained from the review of literature to build and refine instructions and frameworks that solve problems from different stakeholdersBuild the algorithms and methodologies following the previous built instructions and frameworksProductize the algorithms and methodologies to be used for application purpose, such as data publication, service delivery, knowledge update, etc.Quantify the processes and relationships in the physical mechanisms through analysis using data science applications (e.g., R, python, Julia etc.).Engage groups with various interests to capture quantitative and qualitative information to incorporate into decision-making and planning processCreate case studies of big data analytics amongst producers with a focus on enhancing intelligence and building smart management strategies