Dan Berne and Valeria Pesce [from the GFAR blog]

In the week of 20-24 November 2017, GFAR convened a course and symposium on Farmers’ access to data in Centurion, South Africa.

This event is an example of the activities that GFAR wants to promote towards ensuring that communities determine their own needs and their own future, which is one of the key focal areas on which partners in GFAR have agreed to work. Farmers’ awareness of their needs and their rights, in this case data needs and data rights, is key in this process..

GFAR organized the event in collaboration with the Information Training and Outreach Centre for Africa (ITOCA), the Global Open Data for Agriculture and Nutrition (GODAN) initiative and the Technical Centre for Agricultural and Rural Cooperation (CTA).

In the training course, we brought together a group of 16 participants from 14 countries learning, but also discussing passionately, about data that are needed by farmers: data that are actually available and data that are not; who is providing or is expected to provide such data; examples of data services and data-driven apps; as well as about the data that are produced by the farmers and what rights they have on these data. Participants were for the most part farmers’ associations’ representatives (in most cases farmers themselves) while others were actors that have intermediary roles in the farm data value chain, like extension agents, advisers and providers of farm management information systems.

It was on the first day, when discussing Geoffrey Moore’s concept of “crossing the chasm”, that our training coordinator Dan Berne realized that this approach to technology adoption had to be adapted to the realities of smallholder famers. A “chasm” between early adopters (a few enthusiasts) and “pragmatic buyers” (those who want solutions to practical problems and represent the real market) has to be crossed if data and precision agricultural solutions are going to help farmers move from subsistence farming to farming as a business. In reviewing this with the attendees, Dan learned a new word from our South-African colleagues: donga, meaning a deep ravine or chasm.  We were already discovering that some of the standard “rules’ for crossing had to be adjusted and coined the expression “crossing the donga”. And this became the leitmotif of the training: how can data help cross the donga.

Donga was not the only key term we used during this training: our colleague Nike Tinibu from Nigeria proposed to use the distinction between “agripreneurs” and farmers instead of the usual distinction between “subsistence agriculture” (a term which our group didn’t like) and “agribusiness”

So, how can data help cross the donga between farming and becoming an agripreneur? What type of data is most needed? Do farmers have access to it? And how do data rights, data licensing and data ownership frameworks improve or worsen the situation?

The course first reviewed the use of data, especially Big Data, in farming, as digitization of farm activities continues to increase. We identified trends driving the use of data, such as climate change, consolidation among farm industry companies, and the use of mobile devices. We took a look at some of the applications of data with smart sensors in the field and other Internet of Things (IoT) applications.

From a farmer’s perspective we looked at the elements and use of a Farm Management Information System (FMIS). Although it can feel overwhelming, we discussed ways to organize the data around the jobs that farmers need to do. Identifying the jobs that the farmer must do, both in farm management and field operations is key. Farmers should also be clear as to who has what role. For example, who gets to recommend an irrigation plan and who is authorized to order and execute it? Dan encouraged farmers and software service providers to work together to map these data flows.

Then we looked at some of the challenges for farmers in using data, such as accessibility. For example, big data wants to aggregate into a large pattern, while farming is very local and seasonal. Data can be inaccurate, incomplete, tedious to get and expensive to “clean”.

Data rights was certainly high on the list of concerns from participants Dan described the data rights and data ownership framework in the United States, while Nicolene Fourie provided an overview of the general situation in Europe and Africa, focusing on South Africa.

In the United States, data is treated as property, so in principle farmer groups have more say as to how the data is used and transferred. In Europe and Africa, data is treated as intellectual property, currently giving farmers less direct “ownership” over the data they provide. Nicolene went deep into the conceptual framework of data rights, from data governance to the FAIR principles to licenses and data ownership. Participants were invited to describe the situation in their respective countries, in terms of both openness of public data and protection of farmers’ data. One idea that came up in the discussion was to help farmers create a “Minimum Viable Product” (MVP, a product with the minimum necessary features) from their data before they offer it up, so it has some IP protection.

The attendees divided into breakout groups to discuss a few potential controversial scenarios regarding data rights. Common observations that came up in the conversation were that farmers often do not know what happens to the data they provide. Is it being sold or traded for value? How are governments using the data? While providing farm data for the greater good (e.g. determining food scarcity), there needs to be greater clarity. Additionally, there was a strong view that farmers should be fairly compensated for data they provide when that data is used to create commercial value.

Anneliza Collet and Nicolene Fourie presented open data sources on weather, climate and terrain – all vital to farmers. These sources are open to farmers. Some cover limited territory, based on funding and/or scope. Both Anneliza and Nicolene emphasized that data means little unless turned into information, then knowledge, and finally wisdom

Because cellphones – and to some extent smart phones – are now a nearly ubiquitous tool, Stephen Kalyesubula provided an overview of mobile applications. These included applications for farm information management, inputs (seeds, fertilizers, etc.), cashless payments, dairy management, and a general query app for farmers. Agro-Market Day was shown as a good example of a farm app. Stephen emphasized the need for app developers to get to know farmers, understand their problems and design in usability.

A particularly interesting experiment was the value chain exercise that Dan proposed on the fourth day of the course.

Starting from a basic value chain (Farm Input > Farm Operations > Farm Output > Trading > Processing > Retail/ Consumer), participants were invited to identify for each step: the key information needed to make decisions and/or execute a task for each process; if data is readily available, hard to get, or not available at all; what/who are the sources; if the data is very useful, complete, incomplete, confusing, unusable, etc.; what key data is missing and what are the key challenges and opportunities to be addressed.

The group discussions were long and went very much in depth. Some conclusions were common through all the groups:

1. The types of data essential for each step of the value chain were identified in very similar ways. Interestingly, many participants interestingly placed some of them either in a prior step of the value chain (land acquisition, planning) – e.g. for data like climate data, land quality/suitability, market data – or throughout the whole value chain – e.g. price data.
   

   
   

   Essential data were considered (from left to right of the value chain): land quality/suitability, climate data, weather data, crop growth data, all necessary input (especially quality information about it), equipment/machinery (again quality and features), crop calendars and optimal time to harvest, post-harvest good practices, storage facilities; expected yield, market information, information on brokerage, information on market / retailers’ acceptance criteria, traceability requirements.

2. All participants highlighted that the in many cases even when potentially useful data is available (weather data, crop growth, market data in some cases), it doesn’t reach the smallholder farmers (it’s not in a usable format, it’s not in the local language or it uses technical language, it’s not disseminated through the right channels or it’s simply expensive). Everybody also highlighted how the level of availability and accessibility as well as the quality of these data varies widely across countries.
3. Among the key challenges identified by many participants were: first of all, inaccurate data (incorrect or not granular/local enough, or not timely), lack of (objective) quality data on machinery and agricultural input, lack of good FMIS, lack of good record keeping by farmers and, very interesting when thinking about sustainable agriculture, lack of data on long-term implications of land use. Beyond just data issues, participants highlighted broader challenges, like weak collaboration, lack of policies and especially harmonized policies across countries, need for capacity building on data use (for farmers or even better for extension workers).

In the afternoon, based on the results of the value chain analysis, we asked the participants to think proactively about how the situation can be improved and what specifically each actor providing and using data in the value chain (farmers, farmers’ organizations, government, research, private sector / service providers) can do. Additionally, we asked them to tell us what they think they themselves individually can do.

The conclusions were interesting:

1. For actors like government and private service providers some of the actions were identified in the same way by almost all participants. Governments should provide infrastructural data, set data policies in place, provide better information that filters down to the farmers. Service providers should provide services / apps / FMIS that are based on actual farmers’ needs. There were also more precise indications, for instance that governments should have M&E Units to ensure the quality of farm input and provide farm input data and that service providers should make farmers aware of data sources and data ownership terms.

2. 
   

   Very interesting actions were identified for Farmers’ Associations and extension services: first of all, according to all participants, linking smaller farmers’ groups and making farmer groups stronger, to have a stronger common voice. More specifically, farmer groups should use this stronger common voice to negotiate for all farmers, e.g. negotiating with the service providers and buying for the farmers and ensuring farmers own their own data, or presenting a value proposition for investment to banks. As one participant put it, “We cross the donga by being in a group.”
   
   In general, both farmer groups and extension agents should build capacities to collect data and use available data as well as establish the linkage between suppliers and farmers.

3. An interesting action proposed for different actors (FOs as well as Universities or NGOs) is the planning for data collection when participating in projects, in order to mobilize resources for data collection from the beginning and ensure maintenance and continuity (i.e. plan for what happens to data once it’s collected

4. 
   

   Participants also elaborated on what farmersthemselves should do.
   
   Things that were identified as responsibility of the farmer are: participate in training and awareness raising on data issues, take initiative to look for data and be equipped to use the data, be willing to share information among farmers, provide feedback on accuracy of data to the data generators and commit to better record keeping.

5. And finally, participants took some responsibilities upon themselves individually: almost all of them said that, especially after this course, it’s their task to train others: “Now we have become teachers!”

“Now we have become teachers” was definitely the best conclusion that we could hope for after four days of course!

And for this we have to thank our trainers:

• Dan Berne, our training coordinator, from Logom Ag Initiative, United States
• Stephen Kalyesubula, from iLabs@Mak Project of Makerere University, Uganda
• Nicolene Fourie, from the Earth Observation Science & IT of CSIR, South Africa
• Anneliza Collett, from the Land Use and Soil Management Directorate of DAFF, South Africa

and all the incredibly committed and active participants:

We also want to thank the GFAR partners who contributed by recommending or endorsing participants and trainers: the Forum for Agricultural Research in Africa (FARA), the World Farmers’ Organization (WFO), the Confederación de Organizaciones de Productores Familiares del MERCOSUR (COPROFAM), the Young Professionals for Agricultural Development (YPARD), the Kenya National Farmers’ Federation (KENAFF) and the Agricultural Research Council of South Africa (ARC) together with the South Africa Council for Scientific and Industrial Research (CSIR) and Department of Agriculture, Forestry and Fisheries (DAFF).

This training was one step in our work on farmers’ rights to data. Our next steps, again in collaboration with partners in GFAR, are:

• The publication of a White Paper based on the content of the training and the discussions held in the symposium
• The delivery of four recorded webinars in which the trainers will re-purpose their lessons for a broader audience.
• Convening an international expert consultation on legal aspects of open data, as a GFAR Collective Action among partners and invited legal/policy experts, bringing together a broad range of perspectives. This consultation is envisaged to take pace in March 2018.

[All photos: CC BY 4.0, attribute to GFAR – http://www.gfar.net - except for group photo: credit Koketso Ramoonwa and Amanda Chimwaza]