Image: Adobe Stock

Creating FAIR and open agricultural data ecosystems with the Gates Foundation (case study)

Wed Dec 11, 2019
$download_content = get_field('download_content');

Find out more about our project to help agriculture programmes embed good data-management practices and to use data effectively, ethically and responsibly

Find out more about our project to help agriculture programmes embed good data-management practices and to use data effectively, ethically and responsibly

Summary

Data is being used in agriculture to help farmers, researchers and policymakers make more informed decisions. For example, farmers use weather and soil data to decide how and when to fertilise, plant or harvest; and policymakers use data about the impact of interventions in projects aiming to improve pest and disease management when making evidence-based investments and decisions.

In 2018, the ODI spent six months working on a project – funded by the Bill & Melinda Gates Foundation, and led by CABI, a nonprofit inter-governmental development organisation – to help specific agriculture programmes in four geographic areas embed good data-management practices and to use data effectively, ethically and responsibly. The Global Open Data for Agriculture & Nutrition also worked with us on the project.

The ODI team worked closely with project officers from the Gates Foundation and their grantees to help make sure data managed through funded projects is FAIR – that is findable, accessible, interoperable and reusable – and to help them identify how to improve their data-management and data-sharing processes. The project focused on agriculture programmes in four regions: Andhra Pradesh, India; Odisha, India; Ethiopia; and Tanzania.

We worked with some of the in-country grantees and produced tools to help them to develop data inventories and data-sharing agreements, and to deliver data-management plans. This included guidance around FAIR and open data, mechanisms to share data, and information around licences and restrictions.

Our approach involved supporting CABI in developing personas and mapping data ecosystems, to enable an understanding of the broader environment within which data is stewarded, and to help identify patterns, gaps and opportunities. Following interviews with grantees, we analysed data flows, interactions between actors, pain points and learning needs across the regions. Finally, we developed a set of tools to help support Gates Foundation agriculture programmes. 

Key facts and figures 

By helping to make agriculture data more open, we aimed to: 

  • enable more efficient and effective decision-making by stakeholders
  • foster innovation that creates equitable outcomes
  • and drive change through transparency in food production chains. 

We worked collaboratively on the project with:

  • Funder: Bill & Melinda Gates Foundation
  • Lead organisation: CABI, a nonprofit inter-governmental development organisation

The project focused on agriculture programmes in four regions:

  • Andhra Pradesh, India
  • Odisha, India
  • Ethiopia
  • Tanzania

We held three workshops and roundtable discussions with attendees from over 18 organisations.

The tools and guides we produced as part of the project, to help achieve the above goals across this project are:

What was the ODI’s role/what is the story/impact? 

The aim was to make data – produced as part of the Gates Foundation-funded research – more accessible. The Gates Foundation invests heavily in research, but that research is sometimes not widely available, can be lost or may only be accessible to a small group of people. Making the research more accessible could provide wide-ranging benefits, for example, by allowing other researchers to re-use the data and identify trends or correlations with other datasets, or to provide transparency in food production chains. Furthermore, as data sharing is critical to the success of many Gates Foundation programmes – helping to embed good practice and negotiate barriers directly benefits the programmes.

We had already identified a need to provide guidance on how to make agriculture programme data more open. A report by the ODI and the Global Open Data for Agriculture and Nutrition (GODAN) initiative was published in 2017, setting out the results of research carried out into the implementation of open data policies in agriculture programmes. It highlighted the need to make data more open and accessible throughout a programme’s life-cycle, which in turn helped with the scoping and foundation for this project.

We worked alongside CABI to carry out the research phase of this project. CABI conducted interviews with grantees and led engagement in the three countries. The ODI then provided training in data ecosystem mapping to the CABI team, who went on to create data ecosystem maps with grantees at the local level. At the ODI, we produced a summary map, drawing together highlights of the mapping process.

We created a series of personas to explore the challenges around data access. Personas are fictional characters who can help us to understand the learning needs, barriers, motivations, and goals of real people. We developed the personas to help capture the different actors in the ecosystem, and identify their needs and support requirements. These personas were based on interviews and workshops with stakeholders in each of the four regions. 

We created maps of data ecosystems to help put FAIR and open data into context within each region. This involved drawing out relationships and data flows between stakeholders and attempting to solve shared problems. We used visualisation tools to bring these ecosystem maps to life

Alongside CABI, we developed practical learning resources to help different personas overcome some of their challenges. We compiled a register of existing relevant guidance and tools from within the open data and agriculture community, and developed new resources (for example the data inventory and the data-sharing guides) where there were gaps or unmet needs. 

The resources created as part of the project are openly available. The resources we created and other non-peer-reviewed research that the Gates Foundation funds is available on the F1000 Gates Open Research platform. This will support researchers with data management and open access.

The findings helped the Gates Foundation team identify the value of data ecosystems in: creating impact; ensuring ethical and equitable access to data; and maximising the value of the data beyond the initial research team. 

When we presented the common data-access and data-sharing issues from the agriculture programmes (see below), project officers at the Gates Foundation began to see data existing within a dynamic ecosystem. They could see connections between projects and pain points, identify more opportunities to collaborate, and found that grey areas became clearer.

Common data access and data-sharing issues in the Gates Foundation agriculture programmes

  • Local context is a key factor in creating a successful data ecosystem.
  • Data is not discoverable enough.
  • Pain points are mainly institutional, but data quality is still low.
  • Different actors have different learning needs.
  • Monitoring of good data management is not emphasised enough.

“We can’t ignore these data ecosystems anymore. We’ll create problems for grantees and partners if we don’t use them.” [Senior Programme Officer, Gates Foundation]

What was challenging?

The compressed timescale meant we had limited time to understand the various data ecosystems, and to fully understand the various tensions and user needs. While projects necessarily have built-in limitations, this did cause a challenge when working across timezones and organisations.

The timescale also meant there wasn’t much time to execute a full communications/engagement campaign. Our energies were focused on delivering the project and outputs, and while we worked well with partners, and communication was generally good, more comprehensive engagement with the grantees (for example, using a variety of channels and techniques) was difficult within the six-month timeframe. 

The testing and iteration process could have been improved. As the project had a short timescale, there was not time to build in a full testing/iteration process for some of the guidance. While the guides and tools were based on user research and needs – and both grantees and partner organisations helped with reviewing – a comprehensive round of testing, following a soft launch, would have been beneficial.

There wasn’t enough time for comprehensive feedback and follow-up. We met the grantees quite late in the project, and so didn’t have time to take on board feedback or fully evaluate the effectiveness of the tools. For the two-year follow-up project, we’re making sure we share the tools widely, ask for feedback, and follow up with the users once we have incorporated the feedback. 

For future projects, we should aim to effectively match the timescale to the scope of the project, and manage expectations about what is in-scope and what is out-of-scope. Where possible, build in time for pre-project engagement and research; as well as allowing time for mid-project iteration and testing.

What went well/lessons for similar projects

Building a good relationship with the client (Gates Foundation) meant that they then trusted us to run a longer project. A two-year follow-up project resulted from this six-month pilot. 

We strengthened our working relationship with the CABI team. The cross-organisation project work was very positive, both teams were engaged, and we developed working styles that worked across both organisations. CABI is now using ODI tools in some of its other projects, including the Global Burden of Crop Loss project, a project aiming to capture and measure the global impacts of crop losses

The potential impact is wider than agriculture. The project was about working methods and techniques to improve data access and data sharing, and sharing this knowledge with the Gates Foundation so that had a deeper understanding of data management techniques. This learning is applicable across all projects involving data. 

For future projects, we should identify and advise on opportunities to scale and replicate good practice, working methods and re-use of particular tools; ensure learning is scalable and can be applied to other projects within the organisation; and continue to focus on building good relationships to build trust and confidence in ODI advice and delivery.

What have we learned?

Getting advice on engagement and collaboration helps. We went back to the Gates Foundation to ask for advice on how to get more engagement from the grantees. They were happy to help and realised the importance of engagement to the project’s success. At times all it needed was a push from the Gates Foundation.

Working with partners who are domain specialists, particularly for international projects, is essential to creating impact. This allows us to focus on our respective strengths and to build on existing relationships, rather than starting ‘cold’. It is important to have local partners on the ground when working with different countries. We had people from CABI who were regularly in touch with the grantees in the three countries. They kept things going, followed up on discussions and ensured the relationships were nurtured. 

Be aware of other initiatives going on in the same sector. We found there was a lot going on around data access in agriculture, and so there was the risk of duplication of work. We set up calls with other organisations so that we could learn from what they were doing, find out whether what we produce for our project could be of use to the wider community, and ask what we could do to help them. 

For future projects, we recommend carrying out a comprehensive audit of existing guidance, and engaging with people within the relevant sector or communities to get a very clear idea of requirements, current practice, and gaps or unmet needs in existing research and services – to inform the scope of the project.