DECODING THE
DATA ECOSYSTEM
Year in review 2018
2018 ANNUAL REPORT:
Decoding the Data Ecosystem
The CGIAR Platform for Big Data in Agriculture was created in 2017 to coalesce the global, data-driven capabilities of CGIAR to both leverage and shape the digital transformations happening in our sector. It has been a multifaceted and rewarding challenge as we aim to become more skillful data stewards, build new alliances, and develop innovation strategies for digital agriculture.
In 2017 the newly created Platform engaged with a broad range of actors to build an Alliance for a Data Revolution, opening the way for new collaborations to shape the future of digital agriculture in developing economies.
Finding the right mix of data, actors, analysis, and action to build resilient food systems is a complex problem. During 2018, we learned a few things about how to address it:
Decoding the data ecosystem
The CGIAR Platform for Big Data in Agriculture was created in 2017 to coalesce the global, data-driven capabilities of CGIAR to both leverage and shape the digital transformations happening in our sector. It has been a multifaceted and rewarding challenge as we aim to become more skillful data stewards, build new alliances, and develop innovation strategies for digital agriculture.
In 2017 the newly created Platform engaged with a broad range of actors to build an Alliance for a Data Revolution, opening the way for new collaborations to shape the future of digital agriculture in developing economies.
Brian King
Platform Coordinator
Finding the right mix of data, actors, analysis, and action to build resilient food systems is a complex problem. During 2018, we learned a few things about how to address it:
The Platform in the Media in 2018
ORGANIZE
The sheer volume, multiplicity, and variability of big data makes the organization of data challenging. There is a need to find solutions for challenges related to storage, usage, and dissemination of data, to accelerate research and build more data-driven capabilities across CGIAR and the agriculture sector as a whole.
Medha Devare
Organize Module Lead
“Increasingly, we’re not talking just about open data but rather about FAIR, or Findable, Accessible, Interoperable, and Reusable data. We realized that in order to fully reap the benefits of Big Data we needed to provide support to researchers and individual data centers and to clarify such questions as: How are we going to make data findable? What kinds of standards might we use to ensure data interoperability?
The Organize module of the Platform for Big Data in Agriculture is an extension of the work begun through this initiative to support data management efforts in different ways across CGIAR.”
In 2018, the Platform made significant progress in organizing big data and making it FAIR for users in the agricultural sphere and beyond.
- Centers organized 34 trainings and 18 data sprints through Platform support in 2018, with almost 40% of the 1000+ uploaded datasets a direct result of these data sprints.
- We developed guidelines on complying with privacy and ethics considerations in research data management, facilitating regulatory compliance to reduce privacy-related organizational risk.
- The Platform supported Centers to develop policies, trainings and tools in on topics including: semantics for data annotation; metadata entry and compliance tools; licensing, and best practices for privacy/ethics compliance in research data management.
- Datasets and publications discoverable through GARDIAN almost doubled in 2018, with several new geospatial, filtering, and querying/data mining features implemented.
- The Platform supported several key multi-Center working groups within the CGIAR Data and Information Management CoP to facilitate progress and co-learning towards making data sets FAIR and publishing them to open repositories.
GARDIAN: The Ag data search engine set to power new discoveries
A new version of the Global Agricultural Research Data Innovation & Acceleration Network (GARDIAN), a pan-CGIAR data discovery tool, was unveiled at the 2018 Convention in Nairobi. For the first time datasets, publications, and crop varieties across all 15 Centers and 11 gene banks of the CGIAR can be easily found. The Platform helped Centers enhance data management practices, standards, tools, and capacity to deliver FAIR research data, resulting in nearly 100,000 publications and 3,000 well-described datasets in GARDIAN, and is on track to add hundreds more with all Centers observing a common metadata schema in 2019. This is a major step towards making all the scientific knowledge the CGIAR network generates openly discoverable, accessible, and usable.
PUBLICATIONS
DATASETS
NEW ENTRIES IN 2018
Responsible Data Management Guidelines to help protect privacy
AgroFIMS: Your new companion for easy standardization of data collection and description
CG Core: A revised pan-CGIAR metadata schema to enhance inter-operability and discovery
Updated dataset accessibility
The Platform developed/ refined the methods and allocated funding to update the global elevation model dataset SRTM and the Spatial Production Allocation Model (SPAM).
Speeding up data-driven innovation across CGIAR
The Platform is committed to support CGIAR Centers to increase their capacity to use big data tools and speed up data-driven innovation. In 2018, Centers received support to organize seminars, webinars, training sessions on dataset annotation, and ‘data sprints’ to help scientists and research teams improve the curation of their datasets, among other initiatives.
In 2018, as many as 1,174 datasets were published by CGIAR Centers in Dataverse and other public repositories, and are available on GARDIAN.
Thanks to the support of the Platform for Big Data in Agriculture, Centers organized 34 trainings and 18 data sprints and collectively published
datasets in 2018
CONVENE
There are many actors in the agricultural space that are using big data in one way or another. All of them bring value to the table. The Platform – through our annual Convention, Communities of Practice, Youth in Data and other capacity building initiatives – promoted new collaborations during 2018 to encourage the growth of the big data sector in agriculture.
Building on efforts since 2017, our inaugural year, to form the right partnerships and alliances for a data revolution, the Platform, in 2018, further strengthened its position as a global convener around big data in agriculture.
- 2018 Convention: The annual Big Data in Agriculture Convention, co-hosted in Nairobi by ICRAF/ILRI, had over 400 attendees from – 60% of which were non-CGIAR and represented over 150 organizations, institutes, and governments. We had additional participation of over 2500 remote participants via online channels and several million social media views.
- Our Communities of Practice have been particularly active, sharing methodologies and supporting the development of capacity across Centers through webinars, trainings and collaborative products. Members grew to several hundred across the CoPs.
- Shared services: The Platform has enabled access to shared services in support of CGIAR research, including high-resolution satellite imagery (Digital Globe); gridded weather data (IBM and aWhere); high-resolution population data (Oak Ridge National Lab); and secure data management tools (Globus).
- Strategic Partnerships: The Platform, with CIAT and ICRISAT, collaborated with Microsoft Research to begin a design process looking at building a unified Internet of Things architecture and forged a partnership Interactive Voice Response (IVR) provider Viamo.
- Training and capacity building: The Platform launched an online training website providing a space for learning and sharing knowledge. It also produced a number of webinars and an online course to help international agriculture development researchers understand the implications of the General Data Protection Regulation for their work (GDPR).
- A digital transformation for CGIAR: The Platform engaged Accenture Development Partnerships to conduct a high-level assessment of the state of digital strategy to provide CGIAR with key information on global trends in digital agriculture, best practice on data science and how organizations can leverage it for organizational change.
The Convention 2018
The annual Big Data in Agriculture Convention was co-hosted in Nairobi by ICRAF/ILRI. The event had 400 attendees from 150 non-CGIAR companies, 2500 remote participants, and several million social media views.
- Over 400 attendees on location
- Additional 3000 viewers via livestreaming
- 2 million Twitter accounts reached on the first day
- Almost 30 million potential impressions for the whole event
- Top trending topic in Kenya first day
- Stayed in top three during convention
In order to transform agriculture, we must invest in the methods and frameworks that enable connecting the tangible world of food and farming with the digital world of data and analytics.
Doing so requires us to grapple with the incredible complexity of food systems and wrangle with the explosion of global data generated by research and the course of virtually everyone’s social and economic lives.
The domains that intersect with food systems include agriculture, market economics, transport logistics, nutrition, and public policy, to name just a few. And the scale of these domains goes from the level of genes and minerals to plants, animals, crops, and herds to whole farm systems, landscapes, local communities, regional and national governments, and ultimately the global scale.
There are powerful new tools for data collection, management, organization, combination, and analysis to handle this complexity.
We need to consider the whole data ecosystem for food security.
Satellites and telecommunication networks enable researchers to collect high-resolution data across large geographic regions.
Digital Impact Alliance – Using satellite imagery to predict food insecurity in Uganda, Tanzania and Malawi
Dalberg Data Insights – Using satellite imagery and telecom data to monitor agro-pastoralists movements
Sensors, drones, and smartphones can capture granular data at specific locations.
Tanzania Flying Labs – Calibrating satellite data using drones to in Tanzania
Sun Culture – Using soil sensors, hyper-localized weather stations and IoT to optimize solar-powered irrigation systems in Kenya
NuruAI App monitors cassava deseases in Tanzania and Kenya
[blog] FarmBeats: Three major obstacles for IoTs in agriculture
Distributed computing can enable analysis at the edges of the network, giving new power to algorithms for helping us interpret and act in the world.
Gro Intelligence – Harnessing big data to predict shocks to the food system and inform decision makers
[blog] Blockchain in food systems
The Fork – Using blockchain technology to make big data more reliable and useful
Taken together these technologies help us to generate active loops of machine-aided learning, taking in diverse data streams, producing insights, and capturing the data on our actions in the system.
During the 2018 convention, we launched our first Youth in Data initiative. The Platform is actively seeking ways to engage with and support youth interested in digital agriculture and big data solutions for a more sustainable food future.
We invited 10 young Kenyans from the Nairobi area to be our media delegates.
They participated in a media training workshop where they learned how to engage effectively on social media and how to find and report on data stories.
During the convention they interviewed and swapped ideas with experts about the role youths will play in the future of agriculture. Here is a selection of their articles.
CGIAR’s digital transformation
Becoming a data-savvy organization is essential for CGIAR to remain relevant in the Digital Age. Better use of digital data will enable the organization to create robust responses to some of the most pressing challenges of our time, including climate variability, food insecurity and malnutrition, and environmental degradation. Equipping the CGIAR system to better use data in response to these challenges should be at the heart of its nascent digital strategy.
To begin to identify the first steps forward, in 2018 the Platform for Big Data in Agriculture engaged Accenture Development Partnerships to conduct a high-level assessment of the state of digital strategy and how this is reflected in the people, process, and technology investments across the organization.
Below are some of the findings of a high-level assessment of digital strategy across CGIAR. They outline actionable recommendations for the organization to improve its use of data and digital technologies, enabling it to more effectively and efficiently reach the objectives outlined in its strategy.
Technology infrastructure
Forward-looking skills
Digital ecosystem thinking
Data and access management
Agile and digital-savvy leadership
Digital Transformation of CGIAR
Digital Transformation of CGIAR
Technology infrastructure
Investing in the right infrastructure that enables interoperability and collaboration beyond silos will help CGIAR reach its goals more effectively.
Forward-looking skills
Infuse a digital mindset in the workforce by making innovation the focus of training and hiring programs. A forward-looking skills agenda helps the organization develop employees who have the skills needed to get the most out of digital technologies.
Digital ecosystem thinking
The vast range of actors and stakeholders in food security research includes public, private, and non-profit organizations, which might be termed the overall “ecosystem” of the industry or field.
Data and access management
Data mobilization and access management is about driving efficiencies and competitiveness through strong data infrastructure and warehouse capability, combined with the right analytics and communication tools. This will create an environment in which data can be easily found, accessed, combined, and re-used.
Agile and digital-savvy leadership
Leaders that are knowledgeable about the value digital technology are more equipped to address gaps and adjust priorities as needs in the organization shift. They define a clear digital vision, communicate about it effectively, and embed it at every level of the organization.
Partnerships and shared services
Only through partnerships can we reach the scale and agility we need to change food and farming systems. In 2018 the Platform inked a number of agreements with partners and has also enabled access to a series of shared services in support of CGIAR research.
Commercial satellite imagery
In 2018 the Platform conducted several trainings and provided technical assistance to CGIAR researchers to make the most out Digital Globe’s commercial satellite imagery and analytics as well as their processing platform GBDX.
Gridded global weather data
In 2018 the Platform secured research access to validated global weather data from 1979 on a 30-kilometer grid from IBM The Weather Company, which may prove to be a key capability for backtesting predictions about agro-ecologies and farming intensification.
High-resolution population data
The Platform facilitated unfettered access for CGIAR to global gridded population LandScan dataset through securing a subscription with Oak Ridge National Lab.
Secure sharing and transfer of large datasets
The Platform secured a subscription to Globus, a service supported by a global network of universities enabling easy, secure transfer of large datasets and setting of appropriate access permissions for sensitive data
In 2018 CGIAR’s Platforms for Big Data in Agriculture and Gender Research joined up to co-design innovative uses of data science to help bridge the gender divide in data; a key step to ensure women are not left behind in digital agriculture revolution. Leveraging the social science and gender expertise of the Gender Platform, the data science capabilities of the Big Data Platform and in collaboration with the global mobile phone industry association GSMA, the partners designed a research project using mobile network metadata to examine changes in women’s economic empowerment across whole food systems. The first country case will be completed in 2019, and the Platforms are already developing the partnerships and open methods to enable replication of the approach across multiple countries. The opportunity: seeing dynamic trends and changes in women’s economic empowerment from the local to the global in ways that were not even possible just a few years ago. The sector will, as a result, be equipped with new tools for women to drive the digital agriculture revolution.
Internet of Things in agriculture
The Platform, with CIAT and ICRISAT, collaborated with Microsoft Research to begin a design process looking at building a unified Internet of Things architecture that bridges crop improvement, agronomy and natural resource application.
Bringing tailored advice to Malawi farmers
In 2018 the Platform forged a partnership to enhance the existing Airtel/Viamo IVR-based farming advisory service, M’Chikumbe, which already had 726,000 registered farmers in Malawi. The partnership will see Viamo, Michigan State University, Lilongwe University of Agriculture and Natural Resources and CGIAR working together to link leading-edge agricultural analytics with the massive communications channel. The project was selected for a startup grant among over 400 submissions to the Partnership for Green Growth and the Global Goals (P4G), highlighting its potential as a model for large-scale sustainable food system transformations.
Capacity building
Navigating the GDPR
The Platform launched an online training website providing a space for learning and sharing knowledge. In response to the General Data Protection Regulation (GDPR), the Platform launched an online course to help international agriculture for development researchers understand the implications the new regulation for their work.
Access GDPR online course
Webinars
In line with the objective to increase the capacity of learning and knowledge sharing, the Platform launched an online training website where, in 2018, several webinars were held, focusing on digital extension, interactive voice response, privacy guidelines, and more.
The Communities of Practice
Our Communities of Practice aim to leverage technology and new data resources to create broader and deeper impact in programming, as well as to build capacity internally and externally on big data approaches in agriculture. They foster collaboration spaces and opportunities, facilitate connectivity and sharing of methodologies, and support the organization of capacity-building workshops.
Some key collaborative products developed include: design of 100 questions (a harmonized set of common questions in socio-economics research); an inventory of standard vocabularies for livestock research; a CGIAR-wide gap analysis of modeling capabilities; and trainings including: best practices in data management, data mining, agrisemantics, image analysis, and compliance with privacy and ethics guidelines.
Our Communities of Practice aim to leverage technology and new data resources to create broader and deeper impact in programming, as well as to build capacity internally and externally on big data approaches in agriculture. They foster collaboration spaces and opportunities, facilitate connectivity and sharing of methodologies, and support the organization of capacity-building workshops.
Some key collaborative products developed include: design of 100 questions (a harmonized set of common questions in socio-economics research); an inventory of standard vocabularies for livestock research; a CGIAR-wide gap analysis of modeling capabilities; and trainings including: best practices in data management, data mining, agrisemantics, image analysis, and compliance with privacy and ethics guidelines.
Data-Driven Agronomy
Led by 
Mission: Foster innovation in data-driven agronomy and generate actionable information based on data-driven systems to optimize agricultural productivity.
595 members
Read about 2018 activities
Setting standards for digital interventions:
In 2018 the CoP focused on setting standards and validating digital tools for the benchmarking of data-driven agronomy services. A webinar and mini-workshop helped establish a framework.
In 2019 the CoP will explore digital extension services and will involve the entire Platform and other CoPs.
Gap analysis and targeting of interventions:
In 2018 the working group on setting Global baseline for data-driven agronomy interventions collected data from 15 million km² of cropland and 28 million km² of pasture in 75 counties, to create a global picture of mobile/internet technology access and use by smallholder farmers. The results will be published in 2019, providing a roadmap to help NGOs governments, and research organizations better target digital interventions.
Crop Modelling
Led by 
Mission: Foster a better coordinated and more standardized approach to addressing crop modeling challenges in agricultural research.
424 members
Read about 2018 activities
Gap analysis of crop modeling capacity within the CGIAR:
One of the main objectives of the CoP in 2018 was to conduct a CGIAR-wide gap analysis to determine the strengths and weaknesses of each CGIAR Center in terms of modeling capacity, skills and expertise, and to determine how modeling can support other research and outreach activities.
The results from this analysis, expected in 2019, will help the COP design a plan for sharing expertise among CGIAR Centers and collaborators.
The COP has also been working on four reviews documenting modeling activities and potential impacts from CGIAR Centers and partners under each of the four drivers of GEMS: Breeding (G), Environment (E), Crop Management (M), Policy/Socio-Economic (S). These reviews will also be published in 2019.
In 2018, the COP awarded six mini-grants (10-20K) to facilitate the development of key activities, tools, datasets, and model analysis that can facilitate CGIAR’s crop modeling research.
Read the update in the Sep 2018 newsletter →
In 2019 the CoP will also focus on increasing visibility and accessibility of crop research data
Geospatial Data
Led by 
Mission: Facilitate CGIAR’s agricultural research using geospatial data and analysis.
510 members
Read about 2018 activities
Publication of geospatial datasets:
In 2018 the CoP continued the publication of flagship geospatial datasets on GARDIAN. All datasets are annotated with CG Core Metadata Standards, making them F.A.I.R. (findable, accessible, interoperable, and reusable).
Capacity building and knowledge sharing:
The CoP co-organized three training workshops at the 2018 FOSS4G Conference in Dar es Salaam, Tanzania addressing i) drones in agriculture, ii) spatial data handling with R, and iii) R API for aWhere Data.
The CoP also organized a side event on Drones in Agriculture Knowledge Sharing, and made a series of academic presentations throughout the conference.
In 2019, the CoP will finalize and publish a position paper for which a writing workshop will be organized jointly with the Data-Driven Agronomy at the beginning of the year.
2019: Atlas on Adaptation
Another exciting opportunity in 2019 is for the CoP to connect with the Global Commission on Adaptation. The Geospatial CoP suggested to develop an Atlas on adaptation, which would show which areas will be affected by climate change, what kind of agriculture is going on there, how the farming system can be changed, and whether farmers will be able to move to other areas.
Ontologies
Led by 
Mission: Tackle major issues related to semantics for FAIR (Findable, Accessible, Interoperable, Reusable) data in agriculture.
242 members
Read about 2018 activities
Crop ontology and gender
A suit of online tools that enables the submission by the ontology curators of new ontological terms in Crop Ontology was developed for the Roots, Tubers and Bananas (RTB) CRP.
The CoP has also been working with RTB and the CGIAR Gender Platform to include gender-related variables.
AgrO
The Agronomy Ontology (AgrO) has been validated and included in the fieldbook application AgroFIMS. The improvement of AgrO was made possible thanks to the feedback from agronomists who are using the App.
SociO!
The Ontology CoP has been working with the Socio-Economic CoP on the development of a Socio-economic Ontology (SociO!) that allows the standardization of key indicators and will be useful for agronomic survey. Classes and sub-classes of the socio-economic ontology will be identified to classify the concepts that will be extracted from the 100 core questions ‘100Q’ prepared by the Socio-Economic Data CoP.
Fish ontology
The CoP has been working with WorldFish on Fish Ontology.
Webinar series launched
In 2018 the CoP produced its first webinar, with Pier Luigi Buttigieg’s talk on publishing F.A.I.R data for agriculture development, which was given for the International Data Week 2018, Botswana.
PhenoHarmonIS 2018
The PhenoHarmonIS 2018 conference focused on harmonization of germplasm, phenotypic and agronomic data for plants.
Livestock Data for Decisions
Led by 
Mission: Drive informed livestock decision making through the better use of existing data and analyses.
153 members
Read about 2018 activities
Livestock Data for Decisions (LD4D) is led by Supporting Evidence Based Interventions (SEBI) an initiative awarded to The Royal (Dick) School of Veterinary Studies at The University of Edinburgh.
Livestock Fact Check
In 2018, the Livestock Fact Check working group developed and published eight factsheets in an effort to help ensure that information relevant to the livestock sector is robust, up-to-date, and appropriately interpreted. All factsheets were investigated and reviewed by members of the LD4D community. Go to page→
Impact Modelling
The Impact Modelling working group laid the groundwork to working with Bill and Melinda Gates Foundation livestock projects to gain a fuller understanding of their impact and identify ways to better target interventions. Go to page →
Livestock Modellers
Bringing together modellers who work within the livestock sector under the LD4D banner, this group is working to share and consolidate information and approaches to modelling livestock production as an input to partial equilibrium foresight models. Go to page →
LiveGAPS, a set of apps designed to explore productivity gaps has been shared with the community. The LiveGAPS team published the results of their research on the potential for different intervention packages to increase yields and profitability of goat meat production in Ethiopia and India. In 2019, work will begin to explore how to make LiveGAPS more accessible and ensure wider use. Go to page →
Portal of livestock related projects
The CoP has been developing an inventory of current and previous livestock projects, starting with those funded by the Bill and Melinda Gates Foundation (BMGF) and descriptions of associated databases. The portal will expand to include projects by other donors including USAID and DFID. Go to page →
The LD4D Secretariat, hosted by Supporting Evidence Based Interventions (SEBI) is now developing a draft strategy and solutions-oriented roadmap for next 5 years. This includes the development of Livestockdata.org as a platform for the community to share knowledge, data and tools, to be launched later in 2019.
Socio-economic Data
Led by
Mission: Tackle major issues related to semantics for FAIR (Findable, Accessible, Interoperable, Reusable) data in agriculture.
495 members
Read about 2018 activities
100 questions and SociO!
In 2018 the working group on ‘100Q’ identified key indicators common to many farm household surveys in order to make them interoperable to the extent possible. A first draft of the 100 questions was produced in 2018 and a final document is expected in 2019.
This list informed the work carried by the Ontologies CoP and their SociO! initiative on ontologies for socio-economic data.
Blockchain coalition
During the 2017 Big Data Convention a loose coalition was formed of organizations and individuals interested in exploring the possibilities of applying blockchain and blockchain-like technologies to address the seemingly insuperable challenges in agri-food system value chains in some low and middle-income countries. The Blockchain coalition collaborates closely with the Fork and in 2018 the CoP worked on collecting evidence of blockchain use in agri-food systems and whether blockchain works for smallholders.
OIMS: Ontology-Independent Metadata Schema
To make data truly interoperable we need to have machine-readable structural metadata that allows the relatively easy construction of ETL procedures to extract information from any set of data. In 2018 the CoP has developed a prototype of a flexible machine-readable metadata schema, which will be finalized and released in 2019. The CoP has also been interacting with software companies to ensure that interview and survey tools take into consideration information about data standardization and that metadata schema are automatically included.
INSPIRE
The Inspire Challenge encourages the use of big data approaches to advance agricultural research and development. The winners are groundbreaking innovations with real potential for developmental impact, that have mobilized underused or misused data, and demonstrate meaningful partnerships with CGIAR and other sector members.
In 2018 we awarded a total of US$1M
→ INSPIRE CHALLENGE 2018
WINNERS
US$ GRANT / WINNER
→ SCALE-UP 2018
SCALE UP RUNNERS
US$ GRANT / RUNNER
SCALE UP WINNER
US$ GRANT
Our Challenge is getting better at targeting innovation
The Platform studied the Inspire Challenge process of 2017 in light of innovation strategy, examining how the process could better target smallholder farmers with game-changing innovations; the resulting synthesis paper informed some changes to the 2018 application process.
In 2017, only about 8% of submissions targeted smallholder farmers. In 2018 it was over 70%. If we aim to solve agricultural problems faster and at greater scale, we need to measure our progress and change tack as necessary. Generating new ideas is important, and so is finding the ways innovations can take up a place in the larger data ecosystem for food security.
Meet the winners!
2018 scale up winner
#1 SEEING IS BELIEVING
Picture-based advisory services make crop insurance work for smallholder farmers in India
US$100,000
US$250,000
Early results: Pictured-based advisories have improved the correlation between insurance payouts and the damage that farmers experience. Farmers are therefore willing to pay more for a combined advisory-insurance package. If insurers can promote the adoption of risk-preventive measures by farmers, they could potentially reduce insurance premiums.
How does it work?
- Using an App called WheatCam, farmers take pictures of their insured wheat plots with their smartphones.
- Trained local agronomists review the pictures to craft real-time, tailored advisories, which are sent to farmers to mitigate identified risks to their insured crops.
- Improved risk management helps reduce expected insurance payouts – and thus insurance premiums – making crop insurance affordable for smallholder farmers.
2018 scale up winner
#1 SEEING IS BELIEVING
Picture-based advisory services make crop insurance work for smallholder farmers in India
US$100,000
US$250,000
How does it work?
- Using an App called WheatCam, farmers take pictures of their insured wheat plots with their smartphones.
- Trained local agronomists review the pictures to craft real-time, tailored advisories, which are sent to farmers to mitigate identified risks to their insured crops.
- Improved risk management helps reduce expected insurance payouts – and thus insurance premiums – making crop insurance affordable for smallholder farmers.
Early results
Pictured-based advisories have improved the correlation between insurance payouts and the damage that farmers experience. Farmers are therefore willing to pay more for a combined advisory-insurance package. If insurers can promote the adoption of risk-preventive measures by farmers, they could potentially reduce insurance premiums.
2018 scale up runner
#2 AFRICA FARMERS CLUB
A Facebook group and Messenger chatbot to connect and inform livestock farmers in Kenya
US$100,000
US$125,000
Early results: The project has built a Facebook community called Africa Farmers Club and mobile chatbot service. A survey of chatbot users revealed that 92% of the 2,473 respondents that were farming at the time of the survey said the Africa Farmers Club had influenced the way they farmed.
How does it work?
- Machine-learning classifiers are used to analyze and turn unstructured Facebook posts and images into insights.
- Insights are combined with data from ILRI to inform dairy farmers about productivity, markets, and livestock management through a chatbot built on the Messenger platform, which had 30,000 active users as of February 2019.
- Through the chatbot farmers can search over 300,000 farmer-generated questions and answers and get timely, actionable notifications delivered to their phones.
2018 scale up runner
#2 AFRICA FARMERS CLUB
A Facebook group and Messenger chatbot to connect and inform livestock farmers in Kenya
US$100,000
US$125,000
How does it work?
- Machine-learning classifiers are used to analyze and turn unstructured Facebook posts and images into insights.
- Insights are combined with data from ILRI to inform dairy farmers about productivity, markets and livestock management through a chatbot built on the Messenger platform, which had 30,000 active users as of February 2019.
- Through the chatbot farmers can search over 300,000 farmer-generated questions and answers and get timely, actionable notifications delivered to their phones.
Early results
The project has built a Facebook community called Africa Farmers Club and mobile chatbot service. A survey of chatbot users revealed that 92% of the 2,473 respondents that were farming at the time of the survey said the Africa Farmers Club had influenced the way they farmed.
2018 scale up runner
#3 MARPLE DIAGNOSTICS
Mobile and real-time diagnostics for devastating wheat rust in Ethiopia
US$100,000
US$125,000
Early results: The project has developed an affordable, mobile in-field pathogenomics platform called MARPLE (Mobile And Real-time PLant disEase), which provides extremely rapid diagnostics ( < 48H). The data generated by the pilot has been incorporated into national early wheat rust warning systems in Ethiopia, enabling more effective control of disease outbreaks in farmers’ fields.
How does it work?
- Samples of wheat showing symptoms are collected in the field. Their DNA is extracted and sequenced using a mobile MinION sequencer.
- PST individual strains are identified thanks to an automated comparative analysis based on an extensive collection of global yellow rust genomic data. MARPLE provides extremely rapid diagnostics – from field sample to strain level identification within 48 hours.
- Once the strain in question is identified, scientists can pin down the threat to different crop varieties.
2018 scale up runner
#3 MARPLE DIAGNOSTICS
Mobile and real-time diagnostics for devastating wheat rust in Ethiopia
US$100,000
US$125,000
How does it work?
- Samples of wheat showing symptoms are collected in the field. Their DNA is extracted and sequenced using a mobile MinION sequencer.
- PST individual strains are identified thanks to an automated comparative analysis based on an extensive collection of global yellow rust genomic data. MARPLE provides extremely rapid diagnostics – from field sample to strain level identification within 48 hours.
- Once the strain in question is identified, scientists can pin down the threat to different crop varieties.
Early results
The project has developed an affordable, mobile in-field pathogenomics platform called MARPLE (Mobile And Real-time PLant disEase), which provides extremely rapid diagnostics ( < 48H). The data generated by the pilot has been incorporated into national early wheat rust warning systems in Ethiopia, enabling more effective control of disease outbreaks in farmers’ fields.
2018 Start-Up Grant Winner
#4 An Integrated Data Pipeline for Small Fisheries in Timor Leste
US$100,000
The project is building an automated data pipeline to help uncover the hidden contribution of small-scale fisheries in Timor Leste and inform policies to not only help improve their management but also realize their potential for food and nutrition security.
How does it work?
- Observers on the shore and 500 solar-powered geo-tracking devices on the fishing boats take records of the catch.
- The data collected will reveal models of fishing behavior.
- A user-friendly dashboard will present fisheries analytics in a simple, graphical way to help fisheries managers and government partners make informed decisions.
2018 Start-Up Grant Winner
#4 An Integrated Data Pipeline for Small Fisheries in Timor Leste
US$100,000
The project is building an automated data pipeline to help uncover the hidden contribution of small-scale fisheries in Timor Leste and inform policies to not only help improve their management but also realize their potential for food and nutrition security.
How does it work?
- Observers on the shore and 500 solar-powered geo-tracking devices on the fishing boats take records of the catch.
- The data collected will reveal models of fishing behavior.
- A user-friendly dashboard will present fisheries analytics in a simple, graphical way to help fisheries managers and government partners make informed decisions.
2018 Start-Up Grant Winner
#5 CubicA
The hotline for banana growers in Uganda that provides the right info, to the right farmer, at the right time.
US$100,000
The project has developed a farmer hotline called CubicA, which delivers highly tailored information to callers thanks to a machine-learning ‘recommender’ that leverages Viamo’s 321-Service previous caller data, along with other data sources.
How does it work?
- CubicA builds on Viamo’s 321-Service – an Interactive Voice Response (IVR) hotline that provides weather and agricultural information to farmers – but targets more specifically banana growers in Uganda.
- CubicA provides highly tailored information to its callers thanks to a ‘recommender’ system that uses machine learning to leverage the data from calls made by more than a million users since Viamo’s 321-Service since its inception, along with satellite images, weather forecasts, and other data sources.
2018 Start-Up Grant Winner
#5 CubicA
The hotline for banana growers in Uganda that provides the right info, to the right farmer, at the right time.
US$100,000
The project has developed a farmer hotline called CubicA, which delivers highly tailored information to callers thanks to a machine-learning ‘recommender’ that leverages Viamo’s 321-Service previous caller data, along with other data sources.
How does it work?
- CubicA builds on Viamo’s 321-Service – an Interactive Voice Response (IVR) hotline that provides weather and agricultural information to farmers – but targets more specifically banana growers in Uganda.
- CubicA provides highly tailored information to its callers thanks to a ‘recommender’ system that uses machine learning to leverage the data from calls made by more than a million users since Viamo’s 321-Service since its inception, along with satellite images, weather forecasts, and other data sources.
2018 Start-Up Grant Winner
#6 Machine Learning for Smarter Seed Selection in Mexico
US$100,000
The project will develop a system that uses vast amounts of data – including satellite images, performance data, and weather information – and machine learning to identify the ideal mixture of maize seeds that will maximize profit and minimize risk at any given farm in Mexico.
How does it work?
- Using CIMMYT’s data from hundreds of on-farm as well as experimental stations in Mexico and a network of seed companies producing varieties for diverse agro ecologies in the country, BioSense will develop machine learning models that predict the performance of maize seed varieties in a range of different conditions.
- A seed selection algorithm will then identify the mixture of seeds or ‘portfolio’ that represents the optimal trade-off at any given farm.
- Maize farmers in Mexico will know what to plant, and seed companies will know which seeds to market where.
2018 Start-Up Grant Winner
#6 Machine Learning for Smarter Seed Selection in Mexico
US$100,000
The project will develop a system that uses vast amounts of data – including satellite images, performance data, and weather information – and machine learning to identify the ideal mixture of maize seeds that will maximize profit and minimize risk at any given farm in Mexico.
How does it work?
- Using CIMMYT’s data from hundreds of on-farm as well as experimental stations in Mexico and a network of seed companies producing varieties for diverse agro ecologies in the country, BioSense will develop machine learning models that predict the performance of maize seed varieties in a range of different conditions.
- A seed selection algorithm will then identify the mixture of seeds or ‘portfolio’ that represents the optimal trade-off at any given farm.
- Maize farmers in Mexico will know what to plant, and seed companies will know which seeds to market where.
2018 Start-Up Grant Winner
#7 Revealing Informal Food Flows in Vietnam through Free Wifi
US$100,000
The project is offering free Wifi at two traditional markets in Hanoi, Vietnam to monitor food flows at and between the markets and put in place the first pieces of a traceability system that will help tackle food safety issues.
How does it work?
- Wifi routers are installed at two selected markets. All mobile devices present in the perimeter are identified by the routers through their unique MAC (media access control) address, providing data on ‘traffic’ at the markets.
- People who want to use the free Wifi service will need to answer short questionnaires targeting traders, vendors, or consumers. This will provide information on sales across time and space.
- ‘Flash surveys’ carried out at the markets will complete the analysis of food flows.
2018 Start-Up Grant Winner
#7 Revealing Informal Food Flows in Vietnam through Free Wifi
US$100,000
The project is offering free Wifi at two traditional markets in Hanoi, Vietnam to monitor food flows at and between the markets and put in place the first pieces of a traceability system that will help tackle food safety issues.
How does it work?
- Wifi routers are installed at two selected markets. All mobile devices present in the perimeter are identified by the routers through their unique MAC (media access control) address, providing data on ‘traffic’ at the markets.
- People who want to use the free Wifi service will need to answer short questionnaires targeting traders, vendors, or consumers. This will provide information on sales across time and space.
- ‘Flash surveys’ carried out at the markets will complete the analysis of food flows.
2018 Start-Up Grant Winner
#8 Using Commercial Microwave Links (CML) technology to Estimate Rainfalls for Agriculture in Kenya
US$100,000
The project will demonstrate the potential of using recent advances in Commercial Microwave Links (CML) technology to estimate rainfalls in crop production monitoring and help design better rainfall-based index insurance in Kenya.
How does it work?
- Using existing Safaricom CML data, the project will estimate rainfall based on changes observed in the quality of radio signals incurred by rain between emitting and receiving microwave links on Safaricom’s network.
- The project will quantify the reduction of basis risk in index insurance using CML rainfall measurements compared to the alternative methods in an on-going pilot of a bundle of credit and rainfall insurance in Machakos county.
- If conclusive, farmers could benefit from better designed rainfall-based index insurance at a minimal cost since the data is already collected and logged by the cellular operators.
2018 Start-Up Grant Winner
#8 Using Commercial Microwave Links (CML) technology to Estimate Rainfalls for Agriculture in Kenya
US$100,000
The project will demonstrate the potential of using recent advances in Commercial Microwave Links (CML) technology to estimate rainfalls in crop production monitoring and help design better rainfall-based index insurance in Kenya.
How does it work?
- Using existing Safaricom CML data, the project will estimate rainfall based on changes observed in the quality of radio signals incurred by rain between emitting and receiving microwave links on Safaricom’s network.
- The project will quantify the reduction of basis risk in index insurance using CML rainfall measurements compared to the alternative methods in an on-going pilot of a bundle of credit and rainfall insurance in Machakos county.
- If conclusive, farmers could benefit from better designed rainfall-based index insurance at a minimal cost since the data is already collected and logged by the cellular operators.
WHAT’S NEXT
The Convention 2019
Building resilient global food security requires us to navigate a complex net of interactions between the biosphere, economy, technology, and society. Machines and machine-to-machine systems shape and accelerate our social and economic lives, even as climates and ecosystems may seem destined to continue on a tragic path. Our ethical frameworks, our human communities, and our institutions struggle to stay abreast the rate of change. We can no longer consider distinct facets of food security in isolation; we need holistic solutions.
To claim these potential solutions we need trust: in institutions, in firms, in dynamic and expanding human communities, and in the technologies themselves that can help us build the future.
Our 2017 convention was about forming the right partnerships and alliances. In 2018 we focused on the data mobilization and interoperability needed to build an effective data ecosystem. In 2019, we hope you will join us to bring these efforts together in pursuit of holistic solutions to feed the future – byte by byte.
Inspire Challenge 2019
In 2019, too, the Platform will challenge partners, universities, and others to use CGIAR data to create innovative pilot projects that will scale.
Inspire Challenge proposals must be a collaboration between a person or team internal to the CGIAR and an external partner, using this simple partnership matching form. Winning teams will receive $100,000 to put their ideas into practice and will then have 12 months to implement pilots to demonstrate viability. Successful pilots will have the possibility to receive an additional $250,000 of scale-up funding.
A new category, Sensing and Renewing Ecosystems, will be introduced in 2019. Applications will be open from 10 April to 17 June 2019.
Challenge categories:
Inspire Challenge 2019
In 2019, too, the Platform will challenge partners, universities, and others to use CGIAR data to create innovative pilot projects that will scale.
Inspire Challenge proposals must be a collaboration between a person or team internal to the CGIAR and an external partner, using this simple partnership matching form. Winning teams will receive $100,000 to put their ideas into practice and will then have 12 months to implement pilots to demonstrate viability. Successful pilots will have the possibility to receive an additional $250,000 of scale-up funding.
A new category, Sensing and Renewing Ecosystems, will be introduced in 2019. Applications will be open from 10 April to 17 June 2019.
Challenge categories:
Revealing Food Systems
Monitoring Pests & Diseases
Empowering Data-Driven Farming
Sensing and Renewing Ecosystems
Digital Food Systems Evidence Clearinghouse
Building on an effort first developed under the USAID initiative “Digital Development for Feed the Future,” the Platform has been working on a Digital Food Systems Evidence Clearinghouse to showcase all kinds of digital tools – not just big data – that will help practitioners easily identify mature technologies and entry-points for them in agri-food systems. The Clearinghouse will highlight both interventions and evidence on the food system.
An early version of the Clearinghouse was presented at the 2018 Convention in Nairobi.
The Platform is carried out with support from the CGIAR Trust Fund, UKAID and through bilateral funding agreements.
Credits
Project leader: Marianne McDade
Writing & Editing: Marianne McDade & Stefanie Neno
Concept & Web Implementation: Stefanie Neno
