
Door Linda Haartsen

Lely & Rovecom challenge
The WUR Life Sciences hackathon on the 25 & 26th of October is a 32-hour long event to ignite data- and tech-driven collaborations across boundaries. Students, tech entrepreneurs, researchers, and domain experts work jointly on targeted challenges, side by side with representatives of leading Dutch agribusinesses. In this blogpost Lely and Rovecom present their joint challenge.
Feeding Algorithms
Bert Alferink works with software and ICT company Rovecom and has farmers’ blood in his veins. He clearly knows what he is talking about when working on software development for optimizing cattle feeding. “A cow’s diet is fully balanced”, Bert explains, “better than my own diet!”
Rovecom feed management software is used by the feed company specialist to compose a total feed plan together with the farmer. The plan is farm-specific and includes recipes for production groups and (depending on the feeding system) individual cows. It takes the nutritional values of the forages produced at the farm as the starting point and then optimizes a top-up of concentrates, taking into account variables like the composition of all ingredients, availability and market prices. Bert: “milk data are a continuous flow of information on cow health. If we succeed to make it a source of learning we can improve feed plans for optimal health.”


Lisanne de Jong is a mathematician and has worked with Lely for 7 years. Lisanne also grew up on a dairy farm, and this helps in her work. “At present, my work at Lely is to transform data into understandable information that is useful to the farmer and his advisor. It helps that I can visit the home farm - which is managed by my father and younger brother-, and check and test things in practice.” Lisanne is quite excited about the new dataset: “the new dataset will be unique and may really have new insights to offer for feed management of lactating cows.”
Can you help discover the potential of these datasets? Lisanne and Bert, Lely and Rovecom, welcome you to join their challenge!
Feeding algorithms
For this challenge, Lely and Rovecom work together. Lely is a leading international family business providing innovative solutions to dairy farmers. Rovecom is a Dutch software company. Lely milking robots gather data on individual cows, such as milk quantity and quality, indicative for cow health. Rovecom software for feeding management holds data on feed composition and expected intake. Farms that use both Lely robots and Rovecom software have data on the same cows, from both sources. For the hackathon the two companies have compiled a dataset with thousands of cows. This opens an array of possibilities to determine correlations between feeding and cow health, and possibly find new ones!
Newly composed and unique dataset of 5000 cows
This is a first chance to analyse this unique, rich dataset. The dataset is compiled in the weeks prior to the hackathon by specialists. Historical records of the past year (aug18-aug19) of several decades of farms, will be used, meaning thousands of individual cows.
Cows feed on a mix of forages (for example grass or corn silage) with a proportion of concentrates. The feeding data includes quality and quantity data: analysed content of forages and concentrates, the recommended recipe for the mix (TMR) and actual amount of concentrates dispensed. Roughage intake is known at herd level and estimated for individual levels. Concentrate intake, depending on the farm, may be known at individual level.
Goal
In the Lely system farmers and (nutritional) consultants have access to a large amount of output-data (for example milk production, conductivity milk, ruminant activity). Based on this data they can take operation and tactical decisions. A farmer can adjust the amount of concentrate offered to an individual cow when milk production decreases; A consultant can adjust the adviced basic ration when ruminating activity decreases. However these decisions are only based on the output of animals, there are also input factors that have an effect on these output parameters. One of the set of input parameters are nutritional parameters. Expected is that combining in- and output parameters can generate extra information on which farmer and consultant can take better decisions.
Lely and Rovecom want to find together with you an answer on the following questions:
- Does the combination of output data (animal characteristics) and nutritional input data generate extra information
- Can we create, based on algorithms, a software prototype that transfers this data automatically into information.
Rich data
A continuous flow of milk data is a close monitor of cow health. Combined with feed data, this offers a rich source for learning; about animal welfare, nutrition-health relations, and feed composition. It could be helpful for improving cow health, reducing greenhouse gas emissions or ultimately human health via milk composition.
Who are we looking for?
We are looking for aspiring data scientists, animal scientists, vets and bioinformaticians who like to solve a problem using data collected on real farms and animals. We need participants who are able to handle large and complex datasets and analyse data in different ways. You will help us improving animals and animal welfare to make farming more sustainable. You can sign up here.
About Lely
As an international family business in the agricultural sector, we spend every day making farmers’ lives easier with innovative solutions and tailored services. We offer solutions for almost all activities in the cowshed: from milking to cleaning. We provide advice on how to organise a dairy farm smartly with the use of management systems.
About Rovecom
Rovecom is an expert agricultural software developer. FeedExpert is a ration balancing software product for dairy, used by leading feed industries and their clients.
