Latest / Talking Dairy / Precision Dairy Farming Series: Smarter data for everyday decisions | Ep. 6

Transcript

Precision Dairy Farming Series: Smarter data for everyday decisions | Ep. 6 · Talking Dairy

• 0:01
  SPEAKER_01: Welcome to Talking Dairy.
• 0:03
  SPEAKER_01: I'm your host, Jack McGowan from Dairy Insects.
• 0:05
  SPEAKER_01: Hi.
• 0:06
  SPEAKER_01: This episode is part of a special series recorded at the 2025 Precision Dairy Farming Conference in Auto Tahi Christchurch.
• 0:13
  SPEAKER_01: I'm joined by Dr.
• 0:14
  SPEAKER_01: Claudia Kamphaus from Wageningen University in the Netherlands.
• 0:19
  SPEAKER_01: Claudia is a leading researcher in precision dairy farming and data science.
• 0:23
  SPEAKER_01: Her keynote looked at how data collected from sensors, collars, and other technologies can be turned into practical tools that help farmers make better decisions.
• 0:32
  SPEAKER_01: From reducing methane emissions to predicting the impact of management changes before you make them.
• 0:36
  SPEAKER_01: Claudia's work is all about using data to make farming smarter and more sustainable.
• 0:41
  SPEAKER_01: Okay, Claudia.
• 0:42
  SPEAKER_01: Let's start.
• 0:43
  SPEAKER_01: We worked together many years ago, maybe so long ago that Dari and Z was known as Dexel.
• 0:49
  SPEAKER_01: No, it wasn't.
• 0:49
  SPEAKER_01: We were DariNZ.
• 0:50
  SPEAKER_01: It was DariNZ.
• 0:51
  SPEAKER_01: When did you come?
• 0:52
  SPEAKER_01: What brought you to New Zealand then?
• 0:54
  SPEAKER_01: And how did you get to what you're doing today?
• 0:57
  SPEAKER_00: I started my PhD in Utrecht in 2006.
• 1:00
  SPEAKER_00: And then I got an email.
• 1:02
  SPEAKER_00: No, my supervisor told me that they had a new sensor developed in New Zealand.
• 1:08
  SPEAKER_00: I think it was called Censortech.
• 1:09
  SPEAKER_00: CensorTech, I think.
• 1:11
  SPEAKER_00: Where the I was interested in using some of the data to predict mastitis, and that they were very eager to send data over to Utrecht.
• 1:19
  SPEAKER_00: And I said, well, nah, they don't need to send the data, just go there.
• 1:23
  SPEAKER_00: I'll go get it.
• 1:24
  SPEAKER_00: I'll go get it.
• 1:25
  SPEAKER_00: So that was my that was my first New Zealand trip, which I loved.
• 1:31
  SPEAKER_00: Went back to Utrecht, finished my PhD, asked Jenny Diego, hey, maybe there is an opportunity for me here.
• 1:38
  SPEAKER_00: I got an email from Jen L.
• 1:40
  SPEAKER_00: Yeah, there is an opportunity for you here.
• 1:42
  SPEAKER_00: I read that on the computer.
• 1:44
  SPEAKER_00: Went downstairs to my hubby.
• 1:47
  SPEAKER_00: I think we can go to New Zealand.
• 1:49
  SPEAKER_00: Okay.
• 1:49
  SPEAKER_00: So that that was we didn't thought very long about that.
• 1:53
  SPEAKER_00: So that brought me here.
• 1:55
  SPEAKER_00: 31st of December 2010, when we arrived here, and I was very happy about that because I thought, ooh, 31st of December, so it's gonna be heaps of fireworks.
• 2:08
  SPEAKER_00: That was my first disappointment.
• 2:10
  SPEAKER_00: It wasn't that but then okay, so yeah, that brought me here to New Zealand.
• 2:17
  SPEAKER_01: You were with us on the Greenfield project, which was New Zealand's first automatic dairy farm.
• 2:22
  SPEAKER_01: Yeah.
• 2:22
  SPEAKER_01: Then you went back.
• 2:23
  SPEAKER_01: What have you done in the meantime?
• 2:25
  SPEAKER_00: I went back to Wacheninger Business Economics Group with Hank Hoogeveen.
• 2:30
  SPEAKER_00: And I'm not an economist, so that was a bit of a weird place to be.
• 2:34
  SPEAKER_00: But Hank is all about sensor data technology, but then also studying the economic value.
• 2:40
  SPEAKER_00: So I just did a postdoc there and worked on okay, so what does sensor technology need to make it valuable, monetary value for farmers?
• 2:51
  SPEAKER_00: Again, sensor technology was a threat throughout my research.
• 2:55
  SPEAKER_00: After three years finishing my postdoc, I had the choice of continuing doing uni, give people, you know, education.
• 3:04
  SPEAKER_00: No, so I wanted to do I wanted to do the fun stuff research and working with farmers.
• 3:09
  SPEAKER_00: So I had the opportunity to go to Wachenger Livestock Research, which is part of Wacheninger University and research, right?
• 3:16
  SPEAKER_00: Yeah.
• 3:16
  SPEAKER_00: But they don't have any education, so they just do just do research.
• 3:20
  SPEAKER_00: Yeah.
• 3:21
  SPEAKER_00: There was an opportunity to work on big data projects as they refer to, and I had to look it up.
• 3:27
  SPEAKER_00: I didn't know what big data was.
• 3:29
  SPEAKER_00: And that slowly evolved into research using AI, computer vision, slowly evolving to data science because that's just how we continue to do our research.
• 3:41
  SPEAKER_00: We collect increasingly large volumes and we have to rethink how we retrieve information out of that.
• 3:48
  SPEAKER_01: That's the data science.
• 3:50
  SPEAKER_01: Your keynote was about boosting precision dairy farming through data science.
• 3:54
  SPEAKER_01: What does that mean in practical terms?
• 3:57
  SPEAKER_00: It means that those technologies they just generate data and they don't even always generate data that is directly linked to whatever you want to look for.
• 4:07
  SPEAKER_00: So you have to mine that data.
• 4:09
  SPEAKER_00: And you have to work with new data analytical skills to get the information out of it.
• 4:16
  SPEAKER_00: If you use video footage and you collect one terabyte of data a day, how?
• 4:23
  SPEAKER_00: So you have to talk about ICT skills.
• 4:25
  SPEAKER_00: So that's just, I think it's necessary to use data science to eventually get the information out of the data.
• 4:33
  SPEAKER_00: Data's not useful unless you can make sense of it.
• 4:36
  SPEAKER_01: And as you said, farmers already have plenty of data.
• 4:39
  SPEAKER_01: Like farmers in New Zealand have milk records, herd test results, pasture covers, and more.
• 4:44
  SPEAKER_01: Why is it so hard to turn that into useful decisions?
• 4:48
  SPEAKER_00: One of the issues that farmers deal with is yep, they have a lot of information, but the systems that generate these data, whether it's a farm management system or milk recording data, they still have to combine the data themselves and have to look in different apps.
• 5:08
  SPEAKER_00: The farmers have to combine it, yeah.
• 5:10
  SPEAKER_00: They have to look for information in different apps, or in sometimes five different information sources.
• 5:16
  SPEAKER_00: That's that's hard.
• 5:18
  SPEAKER_00: Yeah.
• 5:19
  SPEAKER_00: That's difficult.
• 5:20
  SPEAKER_01: Yeah, you talked about the three V's of big data: volume, velocity, and variety.
• 5:26
  SPEAKER_01: What are these challenges?
• 5:28
  SPEAKER_00: The first one is volume.
• 5:29
  SPEAKER_00: You cannot work no longer in Excel or something because it just has limited records.
• 5:33
  SPEAKER_00: So you cannot use Excel.
• 5:34
  SPEAKER_00: You have to use other analytical tools to deal with these volumes of data.
• 5:40
  SPEAKER_00: You also have to rethink: do we have to save all the data or with just the important information?
• 5:45
  SPEAKER_00: What is the important information then?
• 5:48
  SPEAKER_00: Velocity.
• 5:49
  SPEAKER_00: I remember about six years ago, I had the discussion in Wacheninger that we wanted to collect methane data real time because it's possible that Sniffer is measuring methane continuously.
• 6:03
  SPEAKER_00: Is it important to have that information in real time?
• 6:05
  SPEAKER_00: There were some people talking, no, so why would you have to have real-time information?
• 6:10
  SPEAKER_00: Well, if you skip that opportunity and do nothing with real-time collected data, doesn't make sense to me.
• 6:18
  SPEAKER_00: The question is how to transform that data into information that you can also see real time or see the figures or see the data.
• 6:26
  SPEAKER_00: I'm an animal scientist, I'm trained in relational data.
• 6:29
  SPEAKER_00: You have a cess file with a number of rows and a number of records and numbers in each field.
• 6:34
  SPEAKER_00: But if you use high-dimensional data, sound or video or an image, it's not numbers.
• 6:40
  SPEAKER_00: So you just have to do different things with an image or a video to get the information out of it.
• 6:47
  SPEAKER_00: That's just a variety of technology that we have to learn to work with as a scientist to get the information out of it.
• 6:56
  SPEAKER_01: You mentioned methane measurements, and in your talk you shared an example about methane breeding values.
• 7:02
  SPEAKER_01: Can you talk us about that?
• 7:03
  SPEAKER_01: How did data science make that possible and what could it mean for dairy farmers?
• 7:08
  SPEAKER_00: That example was chosen to show that, well, we had the challenge to reduce methane emissions in the Netherlands.
• 7:15
  SPEAKER_00: We have a deadline of 2030, that's tomorrow.
• 7:18
  SPEAKER_00: If we want to use breeding, we and there was no breeding value in 2018.
• 7:23
  SPEAKER_00: Breeding value is not developed in a year normally, and we didn't even have methane emissions.
• 7:27
  SPEAKER_00: So that point was hey, can we speed things up by rethinking how we're gonna collect methane data and step away from old school data collection using a USB to hey, can we do something else?
• 7:43
  SPEAKER_00: And that is the ICT skill, one of the three domains of what we call data science, that we had to rethink.
• 7:50
  SPEAKER_00: And that allowed us to collect lots of data from individual animals in just four years' time.
• 7:56
  SPEAKER_00: And you need a very smart student, but then you have a then you have a breeding value in five years.
• 8:03
  SPEAKER_00: And that is, I think, that data science can help speed things up.
• 8:07
  SPEAKER_01: So I recall from that example you completely changed the way that you were collecting and processing the data to make it far more real-time.
• 8:17
  SPEAKER_01: Yeah.
• 8:18
  SPEAKER_01: Reduce the amount of data that you were actually holding on to because you were just keeping the good stuff.
• 8:23
  SPEAKER_00: Exactly.
• 8:24
  SPEAKER_00: And I think what was valuable to that one as well is that we offered farmers a way to directly visualize methane emissions, monitoring that you couldn't do anything with it, but hey, I changed my feed.
• 8:37
  SPEAKER_00: Do I see anything?
• 8:39
  SPEAKER_00: So that was also a good way to keep farmers connected to this research.
• 8:44
  SPEAKER_01: For farmers who feel overwhelmed by big data, how much data they're dealing with, the technologies they have now, what's one simple thing they could do to kind of prepare for this more data-driven future?
• 8:58
  SPEAKER_00: Connect to people, talk to people, talk also about your worries, about your fears, and go visit people that are working with it and just ask how are you doing stuff?
• 9:08
  SPEAKER_00: If you have a huge alarm list, how do you deal with that?
• 9:13
  SPEAKER_00: Don't talk to farmers and don't do anything with it.
• 9:16
  SPEAKER_00: Just cherry-pick good farmers and learn from them, talk to them, ask for help.
• 9:22
  SPEAKER_00: I think that's the first thing that you don't try to solve it all by yourself and do nothing, but ask.
• 9:29
  SPEAKER_00: When you're done talking with others, you can buy the most expensive technology and install it in your in your shed or whatever.
• 9:40
  SPEAKER_00: And you don't act on the information that is provided, nothing will change.
• 9:45
  SPEAKER_00: So don't be afraid of acting on information provided by the technology and then keep on monitoring.
• 9:52
  SPEAKER_00: Do you see a change?
• 9:53
  SPEAKER_00: Is the change in the right direction?
• 9:55
  SPEAKER_00: If not, okay.
• 9:56
  SPEAKER_00: But don't do nothing.
• 9:58
  SPEAKER_00: Then you lose money.
• 10:00
  SPEAKER_00: That will not change your performance.
• 10:03
  SPEAKER_00: So learn through trial and error.
• 10:05
  SPEAKER_00: Yeah.
• 10:06
  SPEAKER_00: It's easy for me to say I don't have to milk cows, but I don't have a business that is depending on making good decisions.
• 10:12
  SPEAKER_00: The only thing that I can say, if you don't do anything, it's not worthwhile to install any technology.
• 10:18
  SPEAKER_01: New Zealand dairy farm is uh well known for learning through trial and error and being quite brave.
• 10:24
  SPEAKER_01: So I think that's a good challenge to put out to them.
• 10:27
  SPEAKER_00: I went to the field days as part of this conference.
• 10:30
  SPEAKER_00: I can't remember his name, but the second farmer, he was a banker in Northern Ireland, I think he said.
• 10:36
  SPEAKER_00: And he just decided I'm gonna stop banking, I'm gonna be a farmer.
• 10:40
  SPEAKER_00: I love it.
• 10:41
  SPEAKER_00: Not an option in the Netherlands.
• 10:43
  SPEAKER_00: No way.
• 10:43
  SPEAKER_00: If you do not own a farm, if you're not have a family farm, you will never be a dairy farmer.
• 10:50
  SPEAKER_01: Thank you so much, Claudia, for sharing those insights on data science and dairy farming.
• 10:55
  SPEAKER_01: The key takeaway that I got is that precision dairy isn't just about sensors and data, it's about turning that data into decisions that improve productivity, sustainability, animal welfare.
• 11:06
  SPEAKER_01: From methane reduction to digital twins, which we didn't get a chance to talk about.
• 11:10
  SPEAKER_01: Yeah.
• 11:11
  SPEAKER_01: These tools are shaping the future of farming.
• 11:13
  SPEAKER_01: So start small, get comfortable with the information you already have, and you'll be ready for what's coming.
• 11:18
  SPEAKER_01: Thanks for listening, Marty.
• 11:26
  SPEAKER_01: If you'd like to get connected with DariNZ's latest advice, research, tools, and resources, whether it's reading, scrolling, listening, or in person, you can visit dairynz.co.nz forward slash get-connected.
• 11:40
  SPEAKER_01: And don't forget to hit follow to keep up to date with our latest episodes.
• 11:44
  SPEAKER_01: As always, if you have any feedback on this podcast or have some ideas for future topics or guests, please email us at talkingdairy at dairynz.co.nz.
• 11:54
  SPEAKER_01: Thanks for listening and we'll catch you next time on TalkingDairy.