BIG DATA SOLUTIONS Innovative approaches to overcoming agricultural challenges in developing nations by harnessing the power of analytics June 2015 1 Contents 3 Introduction 4 What is Big Data? 5 Brief Structure 6 CASE STUDY: Providing Dynamic, Targeted Advice to Rural Farmers 12 CASE STUDY: Climate Smart Agricultural Management for Small Holder Farmers 16 CASE STUDY: Ensuring Growth And Prosperity by Facilitating Financial Inclusion in Agriculture 20 CASE STUDIES: The Future of Big Data in Agriculture 22 References 2 Introduction T he food system is fundamental ways. The recent worldwide diffusion of new for human life. It provides technologies, combined with Big Data and Analytics, is providing the opportunity for developing countries energy, nutrition, an income to leap frog some of the intermediate development source for billions of people, and phases providing farmer’s in the developing world is the largest user of the world’s natural with greater access to timely, cost effective, and resources. Yet about 800 million people personally relevant information on best practices, still go to bed hungry every night, and markets, prices, inputs, weather and news of impending disaster. many more suffer from the “hidden hunger” In particular, the penetration of mobile phones, of malnutrition. Furthermore, assuming and the innovative applications of geospatial and current trends continue, the world’s sensing technologies are providing opportunities to growing population is expected to increase use Big Data in support of agriculture initiatives, including the ones funded by the World Bank. by another 2.4 billion people by 2050 , The implementation of these technologies offers while the increasing variability in weather the potential to provide much needed information patterns negatively impacts growing to develop a comprehensive and real-time global seasons and the abilities of farmers to agricultural statistical database that bolster produce and scale their operations to availability of tertiary services for farmers, identify and mitigate the spread of agriculture diseases, accommodate that size of a population. proactively respond to challenges from climate In response, improvements in agriculture change, and deliver real-time information to farmers operations have to be made at both the policy and that can help them to optimize their operations. individual farmer level in order to realize gains This Solution Brief defines what Big Data is in the in efficiency and productivity. Further, these context of the developing world, presents a series improvements need to be realized most in low and of case studies on how Big Data has already been middle income developing countries, where there used to date, and identifies some lessons learned and is currently a lack of traditional infrastructural potential opportunities for the use of Big Data in (farming equipment, financing, distribution systems) supporting the achievement of agricultural outcomes support and an increased vulnerability to food in the developing world. security and climate change. Feeding the growing population in the coming years will require ingenuity and innovation to 1. UN Reports: http://www.un.org/apps/news/story.asp?NewsID=45165#. produce more food, on less land, in more sustainable VY0dxvlVgSV 3 What is Big Data In a Development Context? Smaller and Less Diverse, But Still Powerful. Big Data is normally defined as high-volume, high- the lack of diverse data collection tools, standards velocity and high-variety (the so-called “three Vs” and coordination and infrastructure as well as — volume, velocity and variety) data sets that can high costs associated with the setup of broad-based be analyzed to identify and understand previously data collection systems. However, the breadth of unknown patterns, trends and associations. information even with these smaller, less diverse To date, most attention regarding Big Data has data streams, when combined with advanced been focused on the developed world. However, analytics, can reveal insights that were previously the diffusion of low-cost data producing devices unknown. Furthermore, collected data from passive (e.g., cells phones, internet-enabled cameras), easy systems like cell phones or satellite imagery access to wide-area data collection systems (e.g., reduces the costs of data collection and analysis, satellite systems, drone-based data collection), and promotes an investigative approach to data global access to very-large data systems, and the in contrast to the static nature of merely creating spread of data analytics expertise and low-cost descriptive reports. data analytics packages means that Big Data analytics is a reality for developing countries. In fact, developing countries throughout the Americas, Africa and Asia are increasingly making use of Big Data to help analyze trends and guide policy. It should be noted however, Big Data in the SOLUTION TIME FRAME developing world is typically much smaller and less diverse than in the developed world. In Big Data Solutions the developing world, Big Data is generally in in Development the range of terabytes rather than petabytes or Big Data Near-Term Solutions exabytes, and normally comes from a single data source (e.g., cell phone data, satellite data) due to Big Data Solutions of the Future 4 STAKEHOLDER KEY Business/Agricultural Farmers Policy Makers Support Operations How is the Solutions • Using Big Data to Facilitate Financial Inclusion in Agriculture. Examining the value Brief Organized? that big data brings to incentivize participation of private enterprise in the agricultural economy and share the burden with policy makers of providing financial access to farmers. The creditworthiness This brief presents three major case studies that of a farmer is linked to crop yields and future have been selected to reflect different levels of prices in an extremely volatile commodities market. availability of Big Data solutions. These case Data is collected, used to define, and measure studies (noted below) represent solutions that are individual, community and agricultural market either already in use in developing countries or will risk to encourage participation from multiple be ready for deployment in a development context market and development stakeholders. in the near-medium term (within the next 3 years). Complementing these case studies, this brief also • Using Big Data to Provide Dynamic, contains a section on the future of Big Data in the Targeted Advice to Rural Farmers. Examining context of agriculture – summarizing promising the value of crowdsourcing, data integration, technologies, tools and techniques related to Big analysis and dissemination of information to help Data and associated analytics that could become support individual farming yields and prioritize available to the developing world within the next development and stakeholder operations. Famers 5-10 years. contribute data on which descriptive and predictive Each case focuses on how the information is analytics are performed, resulting in a dynamic relevant for a different set of target stakeholders. feedback loop where the knowledge of one is merged That is, we describe how information can be used at with the knowledge of many and shared, providing the national and regional levels (e.g., city, state, or valuable insights to the entire community. district governments/policy makers), by individual users (e.g., smallholder and commercial farmers • Using Big Data to Support Climate Smart and their communities) and supporting operations Agricultural Management for Smallholder for agriculture (e.g. development organizations and Farmers. Exploration of the use of big data to businesses). understand climatic and weather data sourcing In general, this Solutions Brief is focused less in and analysis along with an assessment of best the particular technology used to collect the data practices for analyzing climate risks using and more on how it has been used and value of the crop-weather interaction models. Practical potential outputs. In particular, the case studies considerations such as the formulation of focus on challenges and issues that the World Bank actionable smart climate management insights for can support, highlighting both what has already farmers and dissemination of relevant information achieved along with some of the less well known are also discussed. stories of Big Data’s use for development. 5 CASE STUDY Dynamic, Targeted Advice to Rural Farmers Providing Dynamic, Targeted Advice to Rural Farmers A gricultural practices, especially in the KEY FINDINGS developing world, have traditionally • For agricultural development objectives to be met, relied on a model of conventional a lack of access to infrastructure and technology in wisdom that is mostly static and often the developing world needs to be overcome with the sub-optimal in the face of the dynamic conditions dynamic collection, analysis and sharing of information faced by most farmers today. Today there is for growth of collective community knowledge. immense pressure on the agricultural community • Big Data and analytics, combined with accommodative to yield output commensurate with the growing technologies can empower policy makers and population’s need to be fed, while maintaining development stakeholders to collect relevant sustainability and ensuring stable livelihoods for information quickly, understand ground realities and farmers. identify problems, prescribe appropriate measures, In this scenario, the use of Big Data and and broadcast relevant information. Analytics can be a key component in empowering • Data is a powerful tool for optimizing efforts from farmers with a collective knowledge base that an operational and policy perspective and mature involves swift and prolific sharing of information systems of data collection greatly facilitate the and insights, filling in the gaps left by conventional advancement of development objectives. wisdom. of a peer network of farmers in remote rural BIG DATA SOLUTIONS communities. These communities have had limited The Grameen Foundation’s Community Knowledge access to traditional infrastructure, and have a Program has managed to achieve this objective in high cost for last-mile manual or infrastructural developing countries like Uganda and Colombia connectivity, making it difficult to set up the (among others) by leveraging technology and data necessary data collection and information broadcast analytics - disseminating and collecting information environment. The peer advisors, or “Community to and from those who have the least access to it. Knowledge Workers,” are in many cases farmers The Community Knowledge Worker (CKW) themselves who are chosen by their peers and program, started in Uganda in 2009, consists respected in their communities. To facilitate communication, the program provides the CKW’s Timeline: Ready to be deployed (Already in place in some with Android phones running an open source, developing countries) custom agricultural information app with an entire suite of data collection tools called TaroWorks (a Decision Makers: Farmers, Policy Makers, product of the Grameen Foundation) behind it. Development Organizations Everything from market pricing, best practices, and disease information is included in the application, Countries Involved: Uganda, Columbia, Ghana and cached for offline access when the phone is off the grid. Data type: Mobile phones, geolocation, crowdsourcing The first time CKWs meet a smallholder farmer, 7 Providing Dynamic, Targeted Advice to Rural Farmers Fig. 1 Heatmap of Chicken Blight in Uganda they register them using the application and facilitate delivery of services that are optimized conduct an in-depth baseline inquiry that contains and targeted to meet specific requirements of the questions about a broad variety of characteristics farming communities. An example of this process that include – socio-demographic information of in action is seen in the facilitation of ‘mSourcing’ farming household (Examples queries: “How many – a process that helps farmer’s associations source children under 11 do you have?”, “Do your children produce from farmers to sell to grocery chains, have shoes? Clothes?”, “What do you use for cooking using the CKW data collection and information fuel?”), farm information (Example queries: “How exchange mechanism to pay farmers quickly, and many of your livestock are currently ill?”, “What to plan production to meet seasonal demand and pesticides are you using?”, “How old are your tree quality standards. crops?”), certification readiness indicators, good agricultural practices adoption metrics and details CURRENT APPLICATIONS on access to finance/services . The collection and One of the most important things to note about analysis of this information provides the Grameen the CKW program is that the communication Foundation the ability to examine the issues facing and sharing of information is bi-directional. A rural smallholder farming communities holistically. rural farmer benefits from the information that is The insights that the data reveal are used by the provided to him by the app on weather forecasts, Grameen Foundation to work with partners and market prices, and best practices. At the same 8 Providing Dynamic, Targeted Advice to Rural Farmers time, the information collected from him/her is DATA COLLECTED BY CKW VOUNTEERS a valuable crowd-sourced data set that is used • Social and Demographic – Characteristics of to perform descriptive and predictive analytics, households, number of people reliant on farm for resulting in a dynamic feedback loop where the livelihood, economic prosperity etc. knowledge of one is merged with the knowledge • Current Farm Condition – Area under cultivation, of many to provide valuable insights to the entire productivity and yield, common pests and diseases, community. The benefits of this bi-directional access to finance and services, etc. sharing of information are augmented by the • Crisis information – Adverse weather events, reports of CKW program’s collaboration with the TECA degradation of soil/water, etc. (Technologies and Practices for Small Agricultural • Operational metrics – Number of people enrolled, time Producers) initiative of the United Nation’s Food taken to complete survey, volunteer productivity. and Agriculture Organization. CKWs share • Best practices adoption and discovery information on agricultural practices from TECA’s knowledge base with farmers and set up TECA exchange groups to review traditional farming and answered over one million questions from practices that farmers have shared. After vetting, over 200,000 thousand farmers in Uganda through successful practices are added to the TECA content the use of the mobile phone technology outlined library for use by the greater community. above (as of 2014)2. This has already resulted Another salient success of the CKW program is in some very interesting and useful results. As in its ability to optimally leverage technology even an example, analysts at Palantir (partners of in the face of many challenges. Even though most the CKW program) used the data collected by of the farmers that this program supports live the program to evaluate the demographics and in communities that are outside of the coverage activities of the most vulnerable farmers (in the of Ugandan cell networks, the phones use GPS lowest 20th percentile of economic prosperity) and satellite signals to record the exact time and were able to drill down on questions relating to location of each query and data input. When the issues concerning farm animals asked between phones return to the grid, all of the data about January and May 2012. This simple result led to the queries are uploaded to a central server. an enhanced ability for different stakeholders This results in the collection of a perfectly geo- (governments, non-profits, individual farmers) to coded dataset that maps out a large proportion detect early outbreaks of disease and parasites of rural farming done in Uganda. The geo-coded in chickens (concentrated in Northern Uganda). and time stamped data is widely regarded as the The coupling of demographic data with farm best data on agriculture in Uganda where the information revealed insights about the problems lack of traditional infrastructure generally makes affecting most vulnerable farming communities – it availability, completeness, reliability and relevance of agricultural data deficient. 2. Palantir Blog Post: https://www.palantir.com/2012/10/grameen-foundation- The Grameen Foundation’s CKW have compiled palantir-partners-for-food-security/ 9 Providing Dynamic, Targeted Advice to Rural Farmers was seen for example that during a period in 2012, STAKEHOLDER INTEREST IN INSIGHTS farmers in southwestern Uganda were experiencing FROM DATA outbreaks of disease in cattle, and that these same • Individual Farmers – Learns best practices and gets farmers commonly did not own clothes and shoes access to services that are at the right place at the (indicating poor economic conditions). Such data- right time driven insights were primarily made possible by the • Non-profits/non-government development partners clean data collection practices of the CKW program. – Look at effectiveness of programs, increase Traditional reporting and surveillance systems in efficiency of delivery of services, develop effective the developing world typically cannot collect the communication and information sharing. data in the same manner as the CKW program • Policy makers/Government Departments – reaches the remotest of farming communities due Identification of targeted issues and food security risk to absence of infrastructure. Further, the use of Big mitigation, analysis and predictability of cost-benefit of Data analytics on the collected data showed how development efforts. economic conditions of rural farming communities can be combined with breakout information on plant and animal diseases, changes in weather rates among other details so that they are always patterns, soil fertility etc., empowering policy aware and can quickly respond to any challenges makers and other development stakeholders not that their workers face. only with early warning and detection signals of • The economic viability assessment and impact oncoming issues, but also with prioritization of tools allow the foundation to identify smallholder development initiatives in real time – such that farmers along metrics of productivity, food they maximize the effect of their efforts to solve security of household, ability to grow, etc. and problems for the people that require it the most. use the insights from the analytics to accurately As the CKW data collection and information identify the needs and corresponding support sharing architecture has matured, the scope that the farmers need (as an example, some of the usage of analytics has also evolved – the farmers only face minor hurdles such as lack of substantial size of the collected data has allowed access to good financing or outdated machinery, Palantir to develop multiple dashboards for the which can be acted upon by the foundation and Grameen Foundation that cover operational its ground partners while other farmers face statistics, economic viability assessments, discovery major challenges such as water contamination, of optimum productivity practices, and good high-dependence of household members, disease agricultural practices adoption rates/program infestations, etc. that need to be escalated to effectiveness. This has transformed the way government stakeholder level for resolution). that foundation carries out its objectives in the • The identification of good agricultural practices agricultural domain to include: and tracking of adoption rates combines • Operational statistics allow the foundation to the knowledge of one with the knowledge monitor data collection/survey speeds and survey of the community (program administrators 10 Providing Dynamic, Targeted Advice to Rural Farmers use the analytics dashboards to identify services and replacing them with families who the most effective solutions to common/ are food insecure; uncommon problems faced by the farmers and • changing the services package to maximize can disseminate the information widely and nutrition outcomes for the population; effectively) while tracking the realities of how • linking families and children who qualify to many farmers are adopting these practices to receive benefits but are not currently enrolled optimize their outreach. in school feeding and healthcare programs combining demographic data collected for FUTURE APPLICATIONS different purposes across different government The use of data has also allowed the Grameen policies; Foundation to look inward and develop strategies The government was able to show that families to improve the CKW program. When the saved on average $300/year on food and that community volunteers share best practices and severe food insecurity was reduced by over 75%. insights disseminated to them through cell phones, This evidence helped make a broader case for the they also have the opportunity to collect data on intervention and is being considered as a model for the adoption of these practices at a later stage, replication by other policy making organizations in giving program managers the ability to examine the country. correlations between specific sets of volunteers The Community Knowledge Program is and workers and high/low levels of best practices a people-involved, relatively low cost model adoption. Insights from the successes and failures that can transform the information-sharing of their efforts allow the foundation to improve the paradigm in some of the most underprivileged program and fine-tune activities. environments in the world. This initiative can be Originally initiated in Uganda, the CKW set up and supported either by governments, aid initiative has now been expanded to Colombia and organizations, or non-governmental organizations Ghana. In Colombia, the Grameen Foundation is with the only significant requirement of effort being working with farmers and government stakeholders the recruitment of community participants and in Salgar, Antioquia, an area that has been hit building of a peer network for farmers. especially hard during the country’s internal conflicts. The government of the state of Antioquia in Colombia used the CKW model to deliver its backyard gardening and entrepreneurship program. Using the information provided, the government helped improve service delivery by • using data to improve targeting: Identifying 25% of the families in the program who were relatively food secure and whose socio-economic conditions could be lifted by providing targeted commercial 11 CASE STUDY Climate Smart Agricultural Management for Small Holder Farmers Climate Smart Agricultural Management for Small Holder Farmers W eather is the single greatest KEY FINDINGS factor influencing a smallholder • Agricultural Output is highly reliant on the climate and farmer’s productivity. Agricultural hence, the livelihood of some of the poorest sections of output and efficiency is extremely the population are vulnerable without smart, dynamic dependent on weather and climate to achieve management of weather effects on farming food production at reasonable costs. Weather • Combining data from multiple sources provides a variability may directly affect production longitudinal view of climate effects and produces levels if farmers are not able to adapt their insights that can be extremely beneficial to farmers systems and practices in time. Around the • Simply collecting the data and generating insights is world, precipitation patterns, maximum and not enough; the use of technology to make sure that minimum temperature cycles, etc. have changed the message reaches the right stakeholders at the right differently in each region and the climate is time is also important. becoming less and less predictable. Traditionally, farmers have used conventional wisdom such as calendar references (when to sow, SOLUTIONS what to sow, when to harvest, etc.). However, due aWhere has empowered smallholder farmers to increasing weather variability, this knowledge is by providing real-time, localized weather constantly being challenged and farmers lack the and agronomic data covering all agricultural information to make the right decisions in a rapidly geographies. This global dataset provides accurate, changing environment. consistent and current observed weather data, It is important for agricultural development as well as forecasted data up to 8 days into the stakeholders/decision makers to extend weather future, and historical weather data for up to 20 smart agricultural management based on big data, years in the past for some regions. The consistency especially to smallholder farmers in developing and availability of this data, along with the ease countries, such that optimal and stable agricultural of access that the platform and the application yields that support food security are ensured along programming interface (API) provide, enables last- with the protection of livelihoods of some of the mile technology partners to pull together insights poorest sections of their population. and convey them to farmers in a cost effective, yet meaningful way. Researchers and commercial practitioners are able to combine weather data with historical information on crop yields to generate Timeline: 1-3 years field-specific agronomic models, as well as weather smart agriculture management recommendations. Decision Makers: Farmers These models and recommendations generate new agricultural intelligence which enhance traditional Countries Involved: Columbia, Ghana agricultural practices and guide farmers on how to mitigate the risks of adverse weather events and Data type: Survey Collected, Mobile, Sensors climate variability. 13 Climate Smart Agricultural Management for Small Holder Farmers CURRENT APPLICATIONS Rice Growers Federation (FEDEARROZ), revealing One crucial component of agricultural intelligence underlying correlation patterns between climate is the ability to create localized recommendations factors and yield variability. or advisories that are matched to growth stage This analysis identified the main growth and models specific to the region in which the farmer limiting climate factors for each region, and allowed operates or even specific to the farmer’s field. In researchers to identify why high or low yields were this effort, it is important to be able to combine observed in a particular crop cycle. This insight, data from disparate sources (e.g. agricultural yield along with local seasonal forecasts, could allow and weather changes) in order to predict future farmers to infer possible future scenarios and agricultural outputs. determine which rice variety and which sowing date An example of the research necessary to inform would result in optimal yield. localized agricultural recommendations was Beyond performing analyses that have the conducted by the Center for Tropical Agriculture potential to provide weather smart agricultural (CIAT). CIAT studied commercial harvest of rice management to the smallholder farmer, it is in Colombia and its relationship with certain key necessary to ensure that these insights are well climatic components from 2007-2013. For each received and implemented. The barriers of access to climate sequence between sowing and harvest information technology infrastructure and poor rates (around 4 months), weather variables were gathered of literacy in developing environments are important from data provided by a climate research institute considerations. Further, the current paradigm of Colombia and cross-referenced with harvest yield of data sharing for weather smart agriculture is information, national rice surveys, and sowing largely unidirectional – with the data providers experiment information provided by the Colombian either supplying the data or derivative insights to their consumers. This model works well in the developed world due to a greater sophistication of infrastructure and availability of tools for automated collection of feedback, but fails to be valuable from both a development and a business perspective in the developing countries. The aWhere platform was developed to overcome the hurdles of providing analytics to smallholders. Its platform consists of open API driven systems that have the ability to track information automatically at a low cost, creating a data-driven feedback loop, which can be used to iteratively improve the sharing of insights and information while maintaining the appropriate balance of abstraction in the context of Fig. 2 Farmers in Malawi, using aWhere and Esoko’s poor technological access and farmer literacy in the Agricultural Data SMS insights developing world. 14 Climate Smart Agricultural Management for Small Holder Farmers A key example of this system’s use is aWhere’s strategies of agricultural communities with partnership with Esoko in Ghana. Esoko has implications for protection of livelihood, poverty developed a robust communications solution on top risk reduction, food security and crisis intervention of aWhere’s weather and agronomic data platform practices. that incorporates mobile messaging, voicemail, Ongoing research shows that farmers using and a farmer call center. aWhere’s API metrics services from Esoko have seen significant for its Esoko customers provide daily baselines on improvements in yield compared to control plots. actual usage of weather information services by the However, the current process for disseminating individual farmers broken down by location and information to rural farmers may hinder its timing. As of 2015, aWhere is providing data to and deployment and democratization in the near term. collecting usage metrics from over 200,000 fields per Although some of the more advanced smallholder day3. farmers are able to benefit from the automatic The primary advantage offered by aWhere as extracts through API’s and delivered through compared to the traditional “ground” meteorological mobile technologies, the majority may still require data collection systems common in developing more traditional dissemination methods like countries lies in the use of Big Data analytics by manuals, reports, or slick sheets delivered in aWhere to combine automatically collected API person during the introductory stages. Moving metrics with satellite data and remote weather forward, aWhere and other providers are not only sensing information, creating a dynamic, accurate considering more ways to extract value from their cheap, customizable and scalable system that data for weather-smart agriculture to smallholders can support the weather management needs of in developing countries but also working on ways smallholder farmers in the most efficient way and means to ensure that useful and sufficiently possible. abstracted information is broadcasted in a cost- effective yet prolific manner. FUTURE APPLICATIONS Promoting weather smart agricultural Combined with contextual information about management has the potential to compensate individual farmers, metrics collected by aWhere for the rapidly declining reliability of traditional on the usage of its API can be analyzed to provide knowledge in the face of increasing climate and insights into the smallholder’s adoption of weather situational variability that many smallholder smart practices from a development policy farmers face today by guiding efficient decision perspective. Gathering statistics on the number of making on what, when, and where to sow, and how farmers subscribing to weather forecasts (broken to manage farms. The impact of better-managed down by gender, location, growing choices), number crops can include increased farm profit and of farmers subscribing to agronomic advice, productivity, higher overall regional food security, number of farmers participating in sustainable and general social stabilization in developing agricultural practices, and the gap between optimal countries. potential yields vs. realized yields can be used to track vulnerabilities in the climate management 3. aWhere Inc. Monthly API metrics report (July 2015) 15 CASE STUDY Facilitating Financial Inclusion in Agriculture Ensuring Growth and Prosperity by Facilitating Financial Inclusion In Agriculture I n order to feed the growing population of KEY FINDINGS the world in the near future, some estimates • Agriculture is perhaps one of the most underserved reveal that overall food production will need to economic sectors. Lack of financial and business grow by over 70% between now and 20504. The services prevent farmers from growing at the rates availability of land for farming is already a strained necessary to feed an increasing and undernourished resource, making any increase in food output world population. increasingly reliant on optimizing farm practices • Big Data from multiple sources and stakeholders can and maximizing the value that can be extracted help provide the essential backbone for the advent of from land. In order to achieve this, both small these support services in the absence of traditional and commercial farming operations all around the infrastructure, providing use to farmers, businesses and world need access to capital, seeds and fertilizers, even policy makers. crop insurance, storage, and distribution. Agriculture in the developing world today is obstructed by a lack of reliable infrastructure. As an sector7. Farmers in the developing world today do not example, the worldwide average number of tractors have the financial means, such as crop insurance, to per 100 sq.km of arable land in 2012 stood at 200, ensure their livelihood and are often forced to arrange whereas in Asia it is closer to 129 and in Africa, it is borrowing at high interest rates because of the lack only 135. Further, despite having close to one-fifth of fair financing options, severely exacerbating the of the world’s (currently utilized) arable land, the strenuous grip of poverty over a large section of the agricultural sector in Africa only accounts for around population. 3% of the world’s fertilizer consumption6. Central in From a policy perspective, moving infrastructure this deficit of infrastructure is the unavailability of towards supporting the needs of the farmers is an financing for farming in the developing world. On especially daunting task that involves large-scale average, agriculture accounts for 35% of the gross investment in machinery, including the expenditure domestic product and employs close to 65% of the of considerable time and effort. Further, micro- labor force in African countries, yet less than 1% of financing and social cooperative efforts work well but commercial lending in Africa goes to the agricultural are often resource strapped, limiting the number of people they can serve. Big Data can play a vital role in easing the process and making financial inclusion in agriculture a possibility. Timeline: 1-3 years Decision Makers: Farmers, Businesses, Policy Makers SOLUTIONS The case of GroVentures (originally started in Ethiopia and currently operating as GroIntelligence Countries Involved: Ethiopia, Multiple East-African Countries in Kenya), showcases how private enterprise can utilize Big Data to participate in the agricultural Data type: Government Authorities, Development Organizations, Crowdsourcing economy and share the burden with policy makers of providing financial access for farmers. 17 Ensuring Growth and Prosperity by Facilitating Financial Inclusion In Agriculture Commercial lending is missing from agriculture relevant for policy and even today is missing because of the difficulty in quantifying the risk information from many East African countries of variability in production. Traditionally, banks since 2011. GroVentures overcame this problem quantify the risk of lending by using multivariate by tracking down and aggregating data from mathematical credit models which predict the a large range of nonprofits, social enterprises, ability of a borrower to repay loans (and determine and developmental agencies working in the the appropriate interest rate). The creditworthiness agricultural space, augmenting this with previously of a farmer is therefore always linked to crop unavailable harvest data from authorities in the yields and future prices in an extremely volatile African Union. This resulted in a consolidated, commodities market. Data can be provide crucial granular and ubiquitous information set for understanding of risk factors involved and all things pertaining to agriculture – from soil encourage participation from multiple market and quality, to historical yield, to correlations between development stakeholders. At the same time, a lack yield patterns and elemental changes. In 35 of of data creates an environment in which providing 54 countries, this meant that, for the first time, support services in agriculture becomes difficult. stakeholders had information on the risk profile of each 100 square kilometers of arable land. The CURRENT APPLICATIONS advent of data and analytics to support agriculture When GroVentures started operations in 2014, had a positive impact on businesses, individual they found that data on historical yields, regional farmers, and policy makers. Businesses are now practices, and general sector related statistics able to effectively allocate capital and loans to was missing. As an example, the United Nations the agricultural sector while using a data-driven Food and Agricultural database (perhaps one approach to fairly evaluate risk and offer affordable of the only long-standing sources of data) only crop-insurance, alleviating the financial strain that contained static, abstracted, high-level information farmers and farm laborers currently experience Did you know? Agriculture accounts Africa would need 3.5 Africa lags behind most for 32% of the GDP and million more tractors to be of the of the world in employs 65% of the labor on par with other areas of fertilizer consumption, force in Africa, yet the world, since it only has accounting for only less than 1% 13 tractors 2.9% of overall of commercial per 100km2 fertilizer lending goes to of arable land consumption agriculture in 2011 18 Ensuring Growth and Prosperity by Facilitating Financial Inclusion In Agriculture in the absence of the data. GroVentures hopes to like GroVentures, such that the pitfalls of a slow, facilitate loans of $25 million to approximately linearized, undeveloped infrastructure are overcome 73,000 farmers in cooperatives and informal by entities that can combine information from farmer organizations via pooled lending by 2015. multiple sources and provide crucial support for This will bring financial inclusion not only to the services that the agricultural industry requires. farm owners, but also to the workers and families attached to the agricultural businesses, affecting directly or indirectly approximately 290,000. Apart from the positive economic impact on a large underserved portion of their population, up-to-date information on market activity and production estimates from the data platforms would also equip Policy Makers to identify vital transportation and infrastructure needs. Furthermore, providing this up-to-date information will help policy makers fulfill immediate objectives while bolstering their ability to better manage grain reserves, and ensure that production shortfalls and market disruptions do not hinder broader population food security. FUTURE APPLICATIONS GroVentures is now embarking on crowdsourcing additional data directly from farmers in order to fill in the gaps that their analytics platform requires. Farmers are beginning to manually self- report information in exchange for mobile airtime credit or access to weather and meteorological services for free. This bi-directional exchange of data and information has an immense potential 4. United Nations Food and Agricultural Organization Report – Global agriculture for optimizing collection practices and monitoring towards 2050 (http://www.fao.org/fileadmin/templates/wsfs/docs/Issues_papers/ HLEF2050_Global_Agriculture.pdf) crucial components of a lending profile, while 5. FAO Statistical Yearbook 2012: Africa Food and Agriculture (Chart 20) http:// pushing out best practices and ensuring their www.fao.org/docrep/018/i3137e/i3137e.pdf 6. United Nations Food and Agricultural Organization Report - Current world fertilizer implementation for a more prosperous outcome for trends and outlook to 2016 (ftp://ftp.fao.org/ag/agp/docs/cwfto16.pdf) all parties involved. 7. World Bank Group International Finance Corporation (IFC) Report - IFC and Policy Makers can support the financial Agri-Finance: Creating Opportunity Where It’s Needed Most (http://www. ifc.org/wps/wcm/connect/44c74a0049585fb1a082b519583b6d16/ inclusion initiatives by providing access to open IFC%2B%2Band%2BAgriFinance%2B-%2BGeneric%2BPresentation%2Bv-4. data and facilitating commercial participants pdf?MOD=AJPERES&CACHEID=44c74a0049585fb1a082b519583b6d16) 19 CASE STUDIES The Future of Big Data in Agriculture The Future of Big Data in Agriculture Data Driven, Cost-Effective Climate Insurance Data Source: Sensors, Cameras Stakeholders: Farmers, Businesses In developed economies all around the world, commercial farmers are turning to the use of a combination of sensors, cameras and probes to generate data and analytics for optimized risk management and cost effective selection of crop insurance. As an example, the Climate Corporation, recently acquired by the Monsanto Company as part of their Big Data and Analytics offerings suite, utilizes field sensors to collect crop yield, soil quality, and weather pattern data, along with farmer reported inputs, to help commercial operations make the best decisions with regards to choosing crop insurance. The technology and data analytics suite quantifies risk factors and combines historical data with predictive intelligence in order to suggest best practices that can mitigate specific risk elements. The platform is currently geared to benefit large-scale operations and requires significant investment in setting up the data collection infrastructure. However, as developing countries leapfrog established technology in favor of cost- effective alternatives (e.g. replacing costly sensor equipment with low cost cameras and software) more and more farmers will be able to incorporate smarter techniques for managing agricultural risks and insuring risks to their livelihood. Mapping with Drones to Feed the World Data Source: Drones, Cameras Stakeholders: Farmers, Policy Makers Remote sensing and monitoring via the use of drones is increasingly becoming a reality for developing countries as the underlying technology becomes cheaper. To take advantage of this increased accessibility that drones will have in the future, Raptor Maps a startup based in the United States, has created a novel platform that uses remote sensing (mainly drones, among other sensors) and associated data analytics to monitor crop and irrigation health, with the objective of managing environmental impact of farming while increasing crop yields. Through a web-interface, the company can provide insights and key statistics directly to farmers and commercial growers. Expanded, this same technology can be repurposed for the remote data collection for policy makers and governments with the associated analytics, resulting in early detection and monitoring of crop and animal disease outbreaks, irrigation and water level control, management of food security, and production deficits, among many others. Using Data and Analytics to Practice ‘Precision Agriculture’ Data Source: Sensors, Drones, Cameras Stakeholders: Farmers ‘Precision Agriculture’ is the collection of real-time data on weather, soil and air quality, crop maturity, and even costs and availability of equipment and labor, followed by predictive analytics that guide smarter dynamic decision making. Traditionally, precision agriculture has been limited only to developed countries and the most large-scale operations due to significant requirements on technology and capital with control centers collecting and processing data in real time and providing direct action items for immediate implementation. This results in farmers making the best decisions with regard to planting, fertilizing, and harvesting crops at any given time. Expensive sensors that need to be placed throughout the fields to measure data points such as temperature and humidity of the soil and surrounding air, yield tracking, and microclimate responses are fast being replaced by insights derived from drones and satellite imagery which are low cost alternatives. Further, in developing countries, the availability and relative cheapness of human capital makes it possible for large-scale manual collection of data – providing low tech alternatives that still achieve the same level of results for precision agriculture. Data-driven precision agriculture - which is already increasing yields by close to 20% and cutting costs by close to 50% in many operations - will be a crucial component that will help the sector sustainably keep up with demand while advancing prosperity for all. 21 References CASE STUDY: Providing Dynamic, Targeted Advice to Rural Farmers 1. Gantt, W. (2014). Grameen Foundation and Learning with Community Knowledge Workers. The Grameen Foundation; 2. Palantir and Grameen Foundation; Kari Knox. (2014, November 25). FOOD FOR THOUGHT: IMPROVING FARMER LIVELIHOODS ON 3 CONTINENTS. Retrieved from The Palantir Blog: https://www.palantir.com/2014/11/food-for-thought- improving-the-lives-of-farmers-on-3-continents/ 3. Palantir and Grameen Foundation; Ari Gesher. (2012, October 12). GRAMEEN FOUNDATION & PALANTIR: PARTNERS FOR FOOD SECURITY. Retrieved from The Palantir Blog: https://www.palantir.com/2012/10/grameen-foundation-palantir- partners-for-food-security/ 4. The Grameen Foundation. (2010, March). COMMUNITY KNOWLEDGE WORKER PILOT REPORT. Retrieved from The Grameen Foundation Publications: http://www.grameenfoundation.org/resource/community-knowledge-worker-pilot-report CASE STUDY: Climate Smart Agricultural Management for Small Holder Farmers 1. Accenture Institute for High Performance. (2012). http://www.accenture.com/Microsites/emerging-markets/Documents/pdf/ Accenture-Esoko-Case-Study-Final.pdf. Retrieved from http://www.accenture.com/Microsites/emerging-markets/Documents/ pdf/Accenture-Esoko-Case-Study-Final.pdf 2. Cariboni, D. (2014, September 30). Colombia rice growers saved from ruin after being told not to plant their crop. Retrieved from The Guardian: http://www.theguardian.com/global-development/2014/sep/30/colombia-rice-growers-climate-change 3. Esoko. (2014, August 28). Will it rain today? Weather Straight to your inbox. Retrieved from Esoko Blogs: https://esoko.com/ will-it-rain-today-the-weather-straight-to-your-inbox/ 4. International Center for Tropical Agriculture (CIAT): http://ciat.cgiar.org/. (2014). Proof of Concept: Big Data for Climate Smart Agriculture - Enhancing & Sustaining Rice Systems for Latin America and the World. World Bank Innovation Challenge Submission Documents. CASE STUDY: Ensuring Growth And Prosperity by Facilitating Financial Inclusion in Agriculture 1. Byrne, C. (2013, September 14). Data-Driven Lending Could Help African Farmers Feed The World. Retrieved from Fast Co.Labs: http://www.fastcolabs.com/3019953/data-driven-lending-could-help-african-farmers-feed-the-world 2. GroIntelligence. (2015, May 29). African Fields to Whole Foods: Potential of Organic Trade in Africa. Retrieved from GroIntelligence Weekly analysis: https://www.gro-intelligence.com/weekly-analysis/Gro-Organic_Potential_May_29_2015.pdf 3. GroIntelligence. (n.d.). Original Analysis by Gro Intelligence. Retrieved from GroIntelligence Website: https://www.gro- intelligence.com/#newsletter 4. GroVentures. (2013, April 12). Big Data as a Tool for Financing African Farmers: Data-Driven Lending. Retrieved from SlideShare Net: http://www.slideshare.net/pafoafri/gro-ventures-cta-presentation-sara-maker 22 References The Future of Big Data in Agriculture 1. Sherrick, B. J., Barry, P. J., Ellinger, P. N., & Schnitkey, G. D. (2004). Factors influencing farmers’ crop insurance decisions. American Journal of Agricultural Economics, 86(1), 103-114. 2. Sharma, R. (2013). Growing the Uses of Drones in Agriculture. Forbes. Available online at:[http://www. forbes. com/sites/ rakeshsharma/2013/11/26/growing-the-use-of-drones-in-agriculture/]. 3. Saari, H., Pellikka, I., Pesonen, L., Tuominen, S., Heikkilä, J., Holmlund, C., ... & Antila, T. (2011, October). Unmanned Aerial Vehicle (UAV) operated spectral camera system for forest and agriculture applications. In SPIE Remote Sensing (pp. 81740H-81740H). International Society for Optics and Photonics. 4. McBratney, A., Whelan, B., Ancev, T., & Bouma, J. (2005). Future directions of precision agriculture. Precision Agriculture, 6(1), 7-23. 5. IBM Research. (n.d.). Precision agriculture: Using predictive weather analytics to feed future generations. Retrieved June 26, 2015, from IBM Research - Articles: http://www.research.ibm.com/articles/precision_agriculture.shtml 6. The Climate Corporation. (n.d.). Climate Insurance. Retrieved June 26, 2015, from The Climate Corportation Website: https:// www.climate.com/mp-federal-crop-insurance/ 7. Subbaraman, N. (2015, May 14). Raptor Maps’ drone technology snags MIT’s $100,000 prize. Retrieved June 30, 2015, from http:// www.betaboston.com/news/2015/05/14/raptor-maps-drone-technology-snags-mits-100k-prize/ 8. Raptor Maps. (n.d.). Technology. Retrieved June 26, 2015, from The Raptor Maps Website: http://raptormaps.com/technology.html 23 Contact This solutions brief is one of several knowledge products delivered through the Innovation Labs and their program in Big Data Analytics. The solution briefs are produced in close cooperation with Global Practices across the World Bank Group. The Innovations Lab sits in the Leadership, Learning, and Innovation vice presidency. Its big data program includes a core program team from the Development Economics Group, the Transport and ICT Global Practice, and Information and Technology Solutions. The purpose of the program is to accelerate the effective use of big data analytics across the organization, and to position the World Bank as a leader in the big data for development community. For additional information about this solutions brief or to find out more about the program, please visit http://bigdata (WBG intranet) or contact Adarsh Desai (adesai@worldbank.org) or Trevor Monroe (tmonroe@worldbank.org). 24