CANBERRA, Australia — Transforming lives and livelihoods through the digital revolution in agriculture was the theme of the 2017 Crawford Fund Annual Conference held in Australia’s Parliament House on August 8.
Research presented at the conference showed the existing capabilities and future opportunities for digital tools to assist in building a food-secure world driven by data.
See more related topics:
Already, delegates said, data is improving our understanding of how the world is producing food and the increasing importance of smallholder farmers in creating variety to support nutritional needs. It is helping in the search for plant gene traits that will grow in various environmental conditions. It can improve the confidence of consumers in buying quality fruit and vegetables. And it is taking mechanisation on farms to the next level, through robotics that improve the quality of food produced and ease the labour burden on an increasingly ageing agricultural workforce. Mobile technology is also already demonstrating the capacity for improved data capture and in-field analysis of plant and animal health.
Yet despite this, research from the McKinsey Global Institute presented at the conference showed that agriculture is lagging behind other sectors in being digital ready. While speakers were keen to highlight the possibilities for big data, there was also an acknowledgement of the gaps that need to be filled and the physical and financial support still needed to produce quality baseline data to support food security and the Sustainable Development Goals.
Devex rounds up the key takeaways from the conference — including the topics that sparked heated debate.
The ‘black art’ of analysis and modelling to fill data gaps in developing countries
In developing countries, any big data — environmental, demographic, health, economic and more — is problematic. A lack of technological capacity, understanding and infrastructure, as well as funding priorities, leads to large gaps in the quality of data between developed and developing countries.
Despite this, attempts are still being made to provide quality data globally. Dr Mario Herrero, chief research scientist with the Commonwealth Scientific and Industrial Research Organisation’s Agriculture Flagship, spoke of the organization’s work to create a global map of field sizes, food variation and nutrients. In many areas, he described the modelling and analysis required as a form of “black arts.”
For developing countries, available data could be incomplete, old, generalized or completely unavailable. But it is important that there are models created for a global picture, he said — requiring researchers and data analysts to make assumptions to fill the gaps.
“It is still important to get the data out there and improved,” Herrero told Devex. “We are now at a point with technology that we can create timeseries data for real analysis of development impact. But we need a starting point.”
Once the data is in the hands of users, including NGOs, improvements can be made and a more accurate picture of the world — including people and the environment — can be formed.
This can not only lead to improved data, but also to improved monitoring of global food production and our ability to meet the SDGs.
Missing: The gender debate
See more Devex coverage of gender data:
With the focus on gender in the SDGs and donor programs, including the Australian aid program, gender has become an important data focus. But at the Crawford Fund conference, it was mainly noted for its absence in big data for agriculture and food security.
“That is not something we have thought about,” Steve Mathews, head of strategy for Gro-Intelligence, told the audience.
Understanding questions about the role of women in agriculture and how programs and policies might increase their access to markets is still a work in progress. Speakers acknowledged it is an area that needs to be better prioritized.
Herrero explained to the audience that CSIRO needs to look further at gender in their work — but the potential is there. A gender focus could be added as data is verified and improved on the ground and within communities. Linking to health datasets can also help to highlight the impact of women farmers within the health metrics of communities.
The question of political mandates for the release of data
Improving the quality and availability of data is a key priority for the data sector in supporting the needs of developing countries now and in the future. But speakers debated the importance of political mandates requiring the release of data.
Profitability is reliant on producing, managing and maintaining high-quality data. This is a better option for improving data in the developing world than relying on politicians to do what they say they will do.—
As Herrero spoke to the audience about the value of data in supporting the monitoring and evaluation of the SDGs, he said that policy drivers requiring the publication of data as part of government investment in the field is an important avenue to improving the quality of data globally.
“If the public paid for the data, it needs to be released,” he told Devex.
But Mathews questioned the reliance on government processes.
“In the United States, we are seeing programs being cut and data obscured from view,” he said. “Any policy that requires politicians to look at data is a problem.”
A commercial model, he suggested, is more sustainable. Profitability is reliant on producing, managing and maintaining high-quality data. This is a better option for improving data in the developing world than relying on politicians to do what they say they will do.
It was a debate with no conclusion, and one that is expected to rumble on if policy continues to hide data from the view of the public and research community.
Open or closed, paid or free?
The data debate continued with speakers discussing why data has been, and still is, kept siloed and locked away.
“Some believe there is a competitive advantage to keeping data siloed,” Mathews said.
Data as a tradable commodity is outweighed by its value as open data, according to Herrero. “Sharing data raises money,” he said. “If I release data openly, I am more likely to be included in new projects and more likely to receive grants.”
He argued that closed data becomes obsolete quickly as open data is used and improved on by public users.
But it was the question of paying people to support the data creation process in developing countries that was hotly debated.
“I’ve been involved in a lot of discussions of getting low-level reliable data, and my answer is to pay for the data,” Mathews said, “People act like that is terrible and something is wrong, but I don’t understand that. You are asking for something of value and people should receive something of value in return.”
He argued that without factoring in a benefit to the data custodians, research projects could die a slow death. Crowdsourcing is too slow and researchers still keep data closely guarded.
But Dr Kim Ritman, Australian chief plant protection officer with the Department of Agriculture and Water Resources, argued that paying for data risked participants telling collectors what they think they want to hear. Dr Lindiwe Majele Sibanda, vice president for country support, policy, and delivery for the Alliance for a Green Revolution in Africa, said she had seen this happen. Honesty is important — acknowledging that you are there to learn will lead to better quality data than paying, she said.
Best practise for open data
An important lesson from the conference was how to make data open. For a sector that is low on the digital scale, opening and sharing data is important for rapid development.
Dr André Laperrière, executive director of the Global Open Data for Agriculture and Nutrition, said that openly available agriculture and food data could end world hunger and provided his tips for creating useable and useful open data to benefit the world.
First, he said, it must be findable. Having data hidden in a corner of the web or worse, locked in a draw, will lead to no one using it, and no one getting benefit from it.
Second, it must be accessible. This means that the barriers to access need to be limited — for example, data should not be closed behind multiple logins; it needs to be in formats that can be read by commonly available programs; and it needs to be accessible by people worldwide regardless of technical, physical or language barriers.
Third, the data needs to make sense. Particularly with environmental data, thousands of images or datasets can be released covering various parts of the globe. But understanding which image or datasets are relevant is a complicated problem. If a wave of data is released with no context, it is unlikely to be used.
Finally, the data must solve a problem. Releasing data for the sake of it helps no one — it just creates another dataset in a catalog gathering dust.
With agriculture and food security still earning its place in the digital landscape, Laperrière urged conference attendees to not just think about data, but to open it.
“Open data is knowledge,” he said. “Big data is for all of us — not just the gods on top of mountains.”
Read more international development news online, and subscribe to The Development Newswire to receive the latest from the world’s leading donors and decision-makers — emailed to you free every business day.