Toekomst met Big Data in AgriFood 7 Juni
Toekomst met Big Data in Agri&Food 7 Juni 2017, Sander Janssen (@wurcgi) & Karin Andeweg Met thema team
Doel: reflectie sessie op strategische kennis vragen rond Big Data in Agri&Food § Start punt: lopende projecten binnen topsector en daarbuiten ● Genereert kennis voor nu en dicht bij toepassing § Strategisch kennisvraag: ● Wat moeten WUR nu ontwikkelen om over 3 -5 jaar in te kunnen zetten? ● Kennis, competenties, skills ● Leidende positie als enabler van technologie 2
Programma, parallele lunch sessie § Presentatie Wageningen UR ● Strategisch thema Big Data ● 3 specifieke voorbeelden § Discussie en reflectie: ● Feedback van iedere deelnemer ● Vragen: ● Hebben we de juiste uitdagingen? ● Inhoudelijke invullingen/aanvullingen? ● Link met topsector activiteiten? 3
4
Big data and related terms Linked data Open data Data revolution Digital agriculture Data science Digital foods Data analytics Digitization 5
Big Data technologies & methodologies § Strategisch thema van Wageningen UR § In afstemming met Min EZ § Methodologisch thema, ondersteunt thematische insteek 6
Big data: ambition 2025 § Big data is business-as-usual ● Big data replaces experimental and one-off data collection for research ● Knowledge is mostly developed on the basis of Big Data § Wageningen UR is trend setting in Big Data analytics and use in the life sciences world wide
Big data in 2018: end of program § Number of leading projects on Big Data, developing: ● Pieces of infrastructure ● Demonstrators ● Consortia § Capacities in hardware, software and orgware to work with and on Big Data § General awareness of what it is, and what it can deliver § Established partnerships with a number of key players 8
Big Data timeline http: //bigdatatimeline. wageningenur. nl/en/ 9
4 V’s of Big Data 10
Our understanding of Big Data http: //dx. doi. org/10. 1016/j. envsoft. 2016. 07. 017 11
Full screen image with title
Access to Big Data: voortgang § In internationale context ● Global Open Data in Agriculture and Nutrition (godan. info) met actieve WUR en NL bijdrage ● FAIR principes gelanceerd § Guidelines voor data governance in samenwerkingsverband § Data governance als obstakel (cross sectoraal) 13
Access to Big Data: plannen 2017 Outcome Activity Big Data collections available for testing in private or public sector setting Interoperable data available for easy linkage with associated good practices Guidelines and best practices on data governance available for stakeholders Agro-Data. Cube project, activities in related projects Scaling out of work with FAIR data points in the Plant Sciences group to other groups Develop guidelines based on inventories in 2016 and distribute, potential follow up of Data. Governance workshop 14
Smart use of Big Data: voortgang § Eerste installaties met Big Data specifieke IT tools (Spark, Hadoop, Cassandra), tests met WUR relevante data § Identificatie en uitwerking van showcases § Start Roadmap ontwikkeling 15
Outcome Prototype of Big Data show case on machine learning for animal genetics Activity Setting up data analytics around animal genetics and experimenting with different machine learning processes Finalised prototype on data Finalising the activities in Ag. MIP Impact visualization on climate impacts across explorer on climate change impacts in a range of data different domains Africa and South Asia Prototype on Big Data analytics for Finding the available data, see the best yield gap analysis in arable agriculture cropping systems (most likely sugarbeets) and setting up analytical soloutions Pilots scoped for large scale With the start of Internet of Food Large innovations with Io. T in agriculture and scale pilots, these pilots will be further food scoped and research/innovation challenges identified 16
Cultural change: voortgang § Seminar series over Big Data @ WUR § Symposium over Data governance § Wageningen Data Competence Centre 17
Cultural Change: plannen 2017 Outcome Activity Awareness of Big data Research Started in 2016 and finishing at Wageningen UR seminar series in 2017 about Big Data at WUR. Seminar are attended well (50 -80 persons). Position of Wageningen UR in national landscape on Big Data established Online materials and presentation of main results of Big Data strategic theme Started in 2016 with Big Data governance workshop, and to be continued in 2017 with another workshop Big Data Dossier made on Wageningen UR website, blog posts added, in 2017 new projects need to be added. 18
Voorbeelden van specifieke projecten § Plant Sciences, Ron Wehrens § Food Sciences, Nicole Koenderink § Animal Sciences, Roel Veerkamp 19
Big Data@Plant Sciences Richard Finkers, Corne Kempenaar, Ron Wehrens
Access to data § FAIRification of data in the plant sciences: ● ● Definition of terms and ontologies Building and maintaining networks with global players Application to current projects Dissemination of knowledge Wilkinson et al. , Nature Scientific Data (2016)
Smart use of data § Novel algorithms for high-throughput phenotyping ● Deep learning, self-organising maps, . . . § “Decent-ware”: professional-level quality software development (software carpentry) ● Application to bioinformatics pipelines
Change in Systems Thinking § Case studies ● Apply results from “Access” and “Smart use” subthemes ● Feedback to research in these subthemes ● Real-life applications (yield gap prediction, . . . )
Betekenisvolle blockchains in Agrifood Nicole Koenderink, Anton Smeenk, Don Willems, Paul Bartels, Jan Top Wageningen Food & Biobased Research
Blockchains in deze presentatie § Technisch: niet interessant, belangrijk dát het werkt § Effect: betrouwbare informatieuitwisseling is mogelijk zonder tussenpartij (financieel en anderszins)
Strategische vraagstelling § Hoe kunnen bedrijven en consumenten ketenrelevante informatie en transactiecontracten vastleggen? ● Betrouwbaar blockchains ● Onveranderlijk ● Begrijpelijke informatie data & context (metadata) semantische blockchains
Semantische blockchains § Blockchains bieden de mogelijkheid om zonder Trusted Third Party transacties af te spreken en vast te leggen § Hiermee wordt ketensamenwerking met willekeurige, onbekende partijen mogelijk § Het expliciet vastleggen van de betekenis van data en contracten is cruciaal om elkaar (automatisch) te begrijpen → semantiek is de sleutel ! § Doel van dit project: de do’s & dont’s rondom semantische blockchains in de agrifood-sector in kaart brengen
Relevantie voor de topsector Agrifood § Blockchains gaan agrifoodketens diepgaand veranderen § Zonder goede meta-informatie: garbage-in-garbage-out § Wageningen UR biedt ● governance: hoe organiseer je blockchains in de keten ● semantiek: hoe zorg je dat er betekenisvolle data wordt gedeeld ● toepassingen in agrifood: ervaringen delen binnen de sector ● samenwerking met aanbieders van Blockchain
Plan van aanpak § Use case ‘Druivenketen’ binnen PPS ‘Blockchains in agrifood’: ● governance: ketenstructuur, rol van partners ● met partners bespreken welke data gedeeld wordt en welke niet ● data + context (met vooraf afgesproken betekenis) § Demonstratie van toegevoegde waarde en overzicht van openstaande issues § Publicatie van resultaten op WUR website en in vakliteratuur Meer informatie via Jan. Top@wur. nl Anton. Smeenk@wur. nl
Big Data in the livestock chain 7 June 2017, Wageningen Roel Veerkamp en Claudia Kamphuis
Big data. . . .
Strategic research question By connecting and combining Big Data and using clever analytical tools can we innovate the sector on the big themes? Management tools Sensor technologies Food chain 32
Examples running projects: 1. Predict cow-individual feed intake in dairy cattle Combine animal nutrition, genetics, cow info, KNMI 2. Sort pigs earlier in life to get homogenous groups Combine on-farm data (litter size, birth information, weight, movements) and genetics 3. Need for management tools for resilient AND efficient production system. Existing sensor data, national data, New technology data (drones) …… 33
Project Discussion with PPS in livestock sector: § Can we predict real-time the norms for nutrient utilisations on a dairy farm? ● Data on grass, cows, youngstock, production, farm, external circumstances, history, soil, manure, silage samples, concentrates … § Can we predict the (expected) survival rate for pigs and poultry for a farm? ● Any data source 34
Discussie § Hebben we de juiste uitdagingen? ● Access to data ● Smart use of data ● Change in systems thinking § Inhoudelijke invullingen/aanvullingen? § Link met topsector activiteiten/disseminatie? ● High Tech to Feed the World ● Andere PPS’en ● Vakbladen/events? 35
Thank you! Sander. janssen@wur. nl @Wurcgi #Big. Data @WUR 36
Challenges for Big data in agrienvironmental domain § More persistent barriers lie in handling the variety and veracity aspects. § Variety: Only bits and pieces of the domain are covered by standards & vocabularies ● Improved semantic interoperability is needed § Veracity: Lack of sufficient, high quality metadata hinders the smooth access to and linkage of data sources. ● Consequence: Lack of trust Lokers et al, 2016, Big Data technologies for use in agro-environmental science, Env. Mod. & Soft: http: //dx. doi. org/10. 1016/j. envsoft. 2016. 07. 017 37
Challenges (2) § Having big data available § Data governance and data sharing across players in the value chain § Sensor integration for crossdisciplinary analysis § Answers looking for questions 38
- Slides: 38