AgroDataCube (KB-33-012-006)

AgroDataCube is an open big data collection for agricultural production in the Netherlands. It brings together many open data sources available in the Netherlands on a consistent basis, and is available through an open source API, via http://AgroDataCube.wur.nl which also envisions activating the community of developers around the Cube to further improve it as a collective public good. As a reference point for the AgroDataCube, agricultural parcels are used and other data sources on soils, weather, height, growth, and water access are connected to these agricultural parcels, to create consistent time series from 2012 to the current year. It aims to become a common reference point for data rich applications in agriculture, where it has been set up with a focus on agricultural production and its environmental performance. The AgroDataCube can thus be offered as data service for application development and it has been used in several hackathons and development sprints as common reference point on data.

A new functional requirement for the AgroDataCube is to have sub-parcel data on 5*5 m grid, which requires a new, better scalable system architecture for the Cube, as the amount of data to be processed rapidly explodes going from 800.000 parcels to millions of grid cells across the Netherlands per year. This project will investigate and implement such an architecture for the AgroDataCube, using commonly available cloud based container services, with the aim to explore and define working procedures for the next generation of computationally intensive projects at WUR.