Big data in agriculture: Between opportunity and solution

CONTEXT

Big data applications in agriculture evolve fast, as more experience, applications, good practices and computational power become available. Actual solutions to real-life problems are scarce. What characterizes the adoption of big data problems to solutions and to what extent is there a match between them?

OBJECTIVE

We aim to assess the conditions of the adoption of big data technologies in agricultural applications, based on the investigation of twelve real-life practical use cases in the precision agriculture and livestock domain.

METHODS

We use a mixed method approach: a case study research around the twelve use cases of Horizon 2020 project CYBELE, varying from precision arable and livestock farming to fishing and food security, and a stakeholder survey (n = 56). Our analysis focuses on four perspectives: (1) the drivers of change that initiated the use cases; (2) the big data characteristics of the problem; (3) the technological maturity level of the solution both at start and end of the project; (4) the stakeholder perspective.

RESULTS AND CONCLUSIONS

Results show that the use cases’ drivers of change are a combination of data-, technology, research- and commercial interests; most have at least a research drive. The big data characteristics (volume, velocity, variety, veracity) are well-represented, with most emphasis on velocity and variety. Technology readiness levels show that the majority of use cases started at experimental or lab environment stage and aims at a technical maturity of real-world small-scale deployment. Stakeholders’ main concern is cost, user friendliness and to embed the solution within their current work practice.

The adoption of better-matching big data solutions is modest. Big data solutions do not work out-of-the-box when changing application domains. Additional technology development is needed for addressing the idiosyncrasies of agricultural applications.

SIGNIFICANCE

We add a practical, empirical assessment of the current status of big data problems and solutions to the existing body of mainly theoretical knowledge. We considered the CYBELE research project as our laboratory for this. Our strength is that we interviewed the use case representatives in person, and that we included the stakeholders’ perspective in our results.

Large-scale deployments need effective interdisciplinary approaches and long-term project horizons to address issues emerging from big data characteristics, and to avoid compartmentalization of agricultural sciences. We need both an engineering perspective – to make things work in practice – and a systems thinking perspective – to offer holistic, integrated solutions.