Research facilitating open data

With research is meant the whole system that produces knowledge and science, including universities, research centres, scientific publications and of course data. Research has its influence on the adoption and use of open data in three important ways: a knowledge reservoir, experimental playground, and objective watchdog. As a knowledge reservoir research provides objective and independent facts and figures, that can assist others in their decision making. While this is the function of research most commonly alluded to and most people will be aware of, the knowledge reservoir on open data in agriculture and nutrition is quite small as it is a new emerging topic.

Some aspects of it have been studied more extensively than others. For example, research on standards and metadata for making data available, and data assimilation in simulation techniques are well known. Other aspects are new and require a lot more attention, like impact evaluations on the benefits of open data, organisational changes caused by open data, and innovative business models for value creation with open data. Impact studies are important to provide public, private and civic sector actors some steer on vital developments in applications – like opening up weather data for farm management and open data on genetic resources for plant breeding.

Impact studies explain what the benefits are and how these benefits come about. They can also demonstrate that the benefits are enjoyed in another part of society than where the costs are. For example, small companies can make farm management applications based on open weather data, but weather agencies need to open up this data. This type of information enables stakeholders to jointly discuss models for benefit sharing. An important research priority on which knowledge is lacking is farmer, small business and consumer participation in the data value in a democratic way, such that each actor keeps control and experiences the benefits of his or her data ownership.

Experimental playground

Research can also offer an experimental playground for testing out developments before to deploy them on large scale. Just having data available as open, will not lead to changes by itself, as data needs to be combined, cleaned out and aggregated till wisdom is achieved (see in the figure Data-Information-Knowledge-Wisdom Pyramid of the Wageningen UR). As often such data combinations and aggregations are new and agriculture and food systems are geographical and temporally specific, research can offer facilities to discover what makes sense (and potentially more importantly what doesn’t).

This can be organised in research projects in which more in depth innovative work is done, like the Yield Gap project or the Food Secure project. Another way is through innovative collaborative sessions like hackathons, Bring Your Own Data workshops, and development sprints (i.e. data and ICT teams working closely together in the development of products). Here researchers both with an agriculture and food background and those with an IT, statistics or data background, can provide expert knowledge to facilitate the process of data value creation during such innovative collaborative sessions. This requires from researchers that they are able to act as brokers in the learning and discovery process of others, while these are using their knowledge and expertise.

An example in which this experimental playground function of research could be instrumental for change is the application of data from remote sensing (also called earth observation). Many detailed satellite images are coming available as open data (for example through the Copernicus program of the European Union). Such data could potentially be used in applications in agriculture such as yield forecasting, irrigation prediction, precision farming. However, it needs experimentation across crops, livestock systems, and farming systems. Observations in highly heterogeneous conditions with small farmer fields, intercropping and agroforestry the application will require a lot tuning and validation, what is the case in for example the CommonSense project (lead by Wageningen UR), which is one of the projects off the GeoData for Agriculture and Water programme (financed by the Netherlands Ministry of Foreign Affairs).

Objective watchdog

Finally, research has a role to play as an objective watchdog to provide wanted and unwanted feedback and common goods around developments of open data in agriculture and nutrition. The watchdog role for open data gives researchers the opportunity to highlight cases of application of open data that are relevant to society at large, while it might not be in the immediate interest of actors to release the data they have, and to highlight issues of privacy, digital divide and gender inclusiveness that could be studied more in detail in the future.

For example, in the Netherlands, there is now a discussion whether the agency admitting pesticides for use in agriculture should release its data on which the admission is based as open, such that it is open for public scrutiny and transparency. With respect to common goods, research can supply data sharing, IT and organisational infrastructure to ensure that all parties can have some level of access. This is also recognised in GODAN, where a Data Ecosystem Working Group is starting to explore infrastructural elements that should be available to all. For agriculture and nutrition, there is an important opportunity here in the sense that such platforms should not become monopolised by any single organisation.

The changing role of researchers

For research to be effective in its various roles, it also needs to change itself. First, it needs to open its own data where possible, so that this is available as a common good. There are movements in this direction with an Open Data Journal for Agricultural Research, many institutional data repositories and developments driven by funders such as the European Open Science Cloud.

Furthermore, research not only needs to lecture and send knowledge, but also listen to the needs to other actors for their needs in research. This makes the evaluation of (potential) impacts of open data across cases in agriculture and nutrition especially important, as this can drive the agenda of research and provide clear indications of obstacles related to a lack of knowledge and understanding. Therefore, in the GODAN Action programme cases on food safety, genetic resources, land data, weather data, nutrition data and food safety were studied with respect to their likely impacts.

In the next years the GODAN Action project will explore several of these cases in more detail, and for each case will work on standards and interoperability for data sharing, specific impact studies on benefits of opening certain types of data, and capacity building on open data of important actors for data-based value creation. The project will thus lead to guidelines on the use of standards, fact sheets and documented cases of impact and many trained individuals, which can also train others. Finally, the programme will produce reproducible methodologies and guidelines for working with open data in new cases. In this way, research will keep channels of communication open during its research process to ensure maximum usefulness of its findings.