The engine room is currently working with Global Open Data for Agriculture and Nutrition (GODAN) to address the issue of how entities in the agricultural sector can balance a practice of ‘open by default’ while still respecting the rights of data subjects and those affected by the release and use of open data. We’ll be working to develop practical solutions for organisations working with agricultural data, using the approach developed through the work of the Responsible Data Forum.

To accurately identify power imbalances and areas where organisations may need to consider how data is opened, we started by looking deeper into who is exchanging data, how they exchange it, and what they are in fact exchanging. Ultimately, we identified 14 archetypes that reflect the main groups of actors producing, managing and using data in this sector and five major ways that these archetypes share data.

We conducted research through publicly available sources to map the actors that participate in the agricultural sector. Then we categorized these actors into different archetypes based on how each archetype exchanges data and with whom. Our use of agricultural archetypes was inspired by Deloitte’s five archetypal actors in the open data ecosystem (see Open Data and Smallholder Food and Nutrition Security). We expanded it to include all actors who are collecting, managing and using data in their work – including those who might open their data in the future. We will detail these archetypes and their relationships in our research in the coming months.

Once we defined the archetypes, we reviewed websites, press releases, publications, and news articles related to each of these entities to work out how they shared data and who they shared it with, distilling information into 5 categories:

• Data shared through the publication of findings, such as the final report from an agricultural research project
• Data translated into a usable form and then sold, such as the use of data collected by commercial UAVs in precision farming
• Data that is analysed and then the results shared, such as the World Bank’s Agriculture & Rural Development Indicators
• Data that is licensed openly, such as through open data portals.
• Data being shared through a partnership, such when a farmer consents to having his data tracked and analysed through an aggregator service, or when two international entities share data as part of work toward a common research goal.

Next, we’ll be identifying some of the key points where actors might usefully think about responsible data issues when opening data. In these initial stages, we’ve already come across some initial questions, which we’ll be diving into over the next few months . We’re interested to hear whether these categories capture the complexity of data use in the agricultural sector – please get in touch.

First, even though business focused entities don’t share data outside of interactions with their clients (generally large-scale or wealthier farmers) they, and social mission-driven organisations, are some of the major entities that directly exchange data with farmers for mutual benefit. For example, for-profit entities provide access to technology such as soil sensors like CropX, telematics for tractor tracking, drones that detect crop stress or potential pest infestations, and data aggregating tools that can track an entire farm. Access to this data can be crucial to help farmers ensure a successful yield. So, do business-focused entities have a responsibility to share this data? Should data about a specific farm be shared externally if it could threaten that farmer’s interests?

Second, despite the fact that a significant amount of existing agricultural data concerns the farmer, very little data flows to them. This issue is magnified where data about smallholder farmers is concerned.  Even when data does flow to the farmer, it is unlikely that they will have the skills and resources to use the data to improve their work/lives without paying for the privilege of doing so. Is there a level of responsibility to share a certain amount of usable data with the farmer whose data is being collected? Consider, for example, the so-called Monsanto Privacy Model. While American farmers came to an agreement with Monsanto over the ownership of their own satellite data, there are few discussions of the same issues in relation to farmers based outside the US. Given our focus on identifying a responsible data framework that is sensitive to the power imbalance between small and large-scale agriculture, it’s important to consider the level of responsibility actors have to share usable data with these subjects.

In the next phase of our research, we’ll be getting in touch with the GODAN and agricultural data community to discuss these topics. Please join us at the ICT4Ag Conference this Friday, June 10 in Washington, D.C. where we’ll be hosting a learning lunch to begin soliciting feedback. Finally, if the project overlaps with your interests or experience, we’d love to hear from you – drop us an email on research[at]theengineroom.org.

This post was co-authored by Kara Kaminski-Killiany and Tom Walker, with input from Zara Rahman and Lindsay Ferris.