Interactive insight in big livestock data
de Jong, A.P.J.
van Waaij, B.D.
The vast amount of livestock (sensor) data collected everyday, offers a huge potential to improve model prediction and therefore decision support for farmers. However, the prediction performance of these models depends highly on the availability and quality of the data. Careful dataset preparation is therefore important. Nowadays, selecting and understanding the data becomes increasingly more difficult as the data grows in size and complexity. In our study we provide data providers within the livestock farming industry with guidelines to describe their data sets more accurately and provide model developers useful techniques to compose a good quality subset of the data to form the base for their model development. This all comes together into our Interactive Visualization tool which allows model developers to explore the data and select, visually or with the use of familiar programming languages, subsets of good quality data for modelling. By using state-ofthe-art distributed computing frameworks, our prototype solution can scale as the size of the data grows to terabytes or even petabytes. We use operational (sensor) data from the Dutch smart dairy farming project to illustrate our solution.
To reference this document use:
Proceedings 9th European Conference on Precision Livestock Farming, Cork, Ireland, 26-29 August 2019