Flies on camera lenses and cable-gnawing mice

Big data has become an integral part of modern farming. Livestock farmer Adrie Vollering and veterinarian and data scientist Miel Hostens discuss the latest technical innovations, screens covered in bird droppings, and how sensors are gradually coming to replace the farmer's natural instincts.

We are in the countryside in Waarder (South Holland) where farmer Vollering - overalls, huge boots and a big grin beneath his hat - leads us to a coffee area in the stable. He takes off his boots and serves us coffee and biscuits. Laughing, Hostens expresses his surprise at the coffee creamer. 'Why do I always get milk in a plastic cup every time I visit a farm!' Despite his job as a data scientist at the Faculty of Veterinary Medicine, he is anything but a typical nerdy scientist who only feels comfortable with graphs and tables. Hostens feels right at home on Vollering's farm. The two laugh a lot and can almost finish each other's sentences.

Favourite app

'We can't live without data anymore these days', Vollering explains as he takes his smartphone out of his breast pocket. He shows us his favourite app, which offers real-time insight into 'his ladies'' data. 'I check the app a few times a day and get push notifications every now and then; for example, if a cow is more or less active than usual.' Hostens nods approvingly. 'It's amazing how much data a farm like this generates. The number of steps each cow takes, the number of rumination movements, the cow's location in the barn, the composition of the milk. But also: the type and amount of feed the animals receive, the quality of the grass in the field, and so on. Vollering adds: 'We were really flooded with data when we started measuring. I realised we needed to find a way to apply it all effectively.'

Morning routine

Vollering used to get up at five in the morning to milk the cows. 'I'd do that till 7.30, and then go take care of the calves. Then it was time for breakfast, and I'd finally have time for the cows that needed extra attention. That could range from dealing with an udder infection to a cow that had to be artificially inseminated. These days, his morning routine is very different. 'Four milking robots do all the milking, which saves me at least six hours of work a day. I arrive at the barn at 6.30 and can immediately focus on the cows that need extra attention.' Bleep! A push notification: cow 242 needs to be inseminated. 'That's the first thing I'll be doing after you leave. Let's see where 242 is hanging out.' Vollering locates the cow with two simple taps in the application. 'Ah, she's in the milking robot now, in the front right corner of the barn.'

Four milking robots milk the cows, saving me at least six hours of work a day

A trained eye 

Vollering opens another application to show us how much milk 242 is currently producing, the amount she usually produces and the protein and fat content of her milk. The system also detects diseases like mastitis. 'I used to have to taste the milk for that. If it tasted salty, I'd know something was wrong. These days, Vollering relies on his sensors, although they do have their limitations. 'The data doesn't always give you a clear picture. For example, I might get a notification that one of the cows is potentially fertile, but I still have to check myself to be sure. Occasionally, Vollering will even pick up on signals before the sensors do. 'I'm crazy about my cows. I was walking through the barn this morning and noticed a cow kept shifting its weight from one leg to the other a little more than usual. It calved just half an hour later.' Hostens nods approvingly again. 'That's because you've got such a trained eye. Ideally, we would like to be able to use sensors to predict when a cow is due to calve. That's a crucial moment in which farmers can prevent the calf from dying at birth. If there was an app that could notify you that a calf was about to be born in the middle of the night, Adrie would immediately jump out of bed. It's up to science to develop those kinds of prediction models.' As Hostens explains, the faculty would actually like to take things a step further. 'Big data will also allow us to contribute to social issues, such as improving the sustainability of livestock farming.'

Big data will also allow us to contribute to social issues, such as improving the sustainability of livestock farming

A world of data 

As we notice on our tour of the barn, data measurement in this type of environment isn't necessarily easy. 'You can't just hang up sensors everywhere', Hostens explains. 'Flies will land on the camera lens, mice gnaw through your sensor cables, and the screen and keyboard inevitably get covered in dust, mud and bird droppings.' And those are just the practical challenges. 'There's so much data to measure. The question is: what do we do with all that information? That's going to be a real challenge for veterinarians moving forward.' Hostens is fascinated by the challenges and potential of farming data. 'I want to pass that sense of fascination on to my students. Farmers and there cows represent a whole world of data. I want to prepare new vets to analyse that data objectively and use it to identify problems and solutions.'

This is an article from: