Data collection for dairy herds: What to consider
Developing cow technology in response to the trend for fewer dairy farms milking more cows, and ongoing labour shortages, has to be more than just a data generating system.
Capturing raw data for the sake of it is not useful – data needs to be turned into information that is beneficial to the dairy farmer, says Scotland’s Rural College’s (SRUC’s) Dr Mazdak Salavati of Crichton Royal Farm, Dumfries.
“We focus on the function: what is the data going to do?
“Raw data is worthless; the value lies in the information you extract from it,” he says, adding that the goal is to make tools that help the decision-making process.
See also: How to use data wisely to inform your dairy business choices
Data-led dairying
Things to consider about the use of data in dairying include:
- There is a difference between data and information
- Data need to be turned into useful information for action
- Data storage for analysis is a big problem because it involves huge amounts of recordings
- Poor farm internet connections affect data upload
- Could dairy farmers lose milk contracts because of a lack of connectivity?
- Cyber security of on-farm data
- Who owns the raw data?
- What is the end goal? (Not to replace humans, but help them)
- Results are commonly presented in red or green, but 8% of males and 0.5% of females are colour blind (colourblindawareness.org)
- Researchers are teaching technology about dairy cows that might be ready to be of practical use in a couple of years
- What is the carbon footprint of camera-sourced data?
Source: Hannah Dairy Research Foundation conference
Adding value
Farmers do not have time to work out what data means – they need it to be analysed, and action suggested, agrees UK Agri-Tech Centre’s director of farms, Dr Rob Morrison.
“Farmers want decisions, not data,” he says.
However, he points out that whoever is adding value to data needs to be rewarded for storing, processing and enhancing it.
Replacing wearables with cameras
Smart camera technology to monitor animal health could replace activity collars, pedometers and tail sensors, and be “less invasive”, says Dr Christos Tachtatzis of Strathclyde University.
But such technology is still in its early stages because current equipment bought to track cow behaviour lacks the human’s ability to identify cows in the background, or those blocked by others in a group, he says.
Another issue is the requirement for a big server to handle the amount of data produced over 24 hours, seven days a week.
“For 16 cameras, this is 55,296GB a month, which is a lot of data over a year,” he points out.
This is where farm connectivity – or lack of it – becomes a problem for data collection and processing, adds Rob.
He says that UK Agri-Tech is working with government to challenge internet providers to get farms connected to networks capable of handling large volumes of raw data.
Cyber security
Furthermore, he says anyone looking to sell technology to farmers needs to guarantee that it will work with connectivity such as wi-fi or satellites.
However, he also raises the threat of cyber security given the number of devices on farm that can now connect to the internet.
Raw data ownership is another concern. Clarifying ownership and access for both service and software provider, as well as the farmers who use it, is important, says Mazdak.
He points out this is where checking written terms and conditions plays a key role before a farmer signs up to a specific product or service.
Farmer wellbeing
As the potential for on-farm technology grows, a social science project is investigating its impact on a farmer’s wellbeing.
Dr Vanesa Fuertes, based at the University of the West of Scotland, Paisley, says that surveys show mental health is already problematic among dairy farmers.
While fitting cows with wearables (such as collars and pedometers) can decrease the daily workload and benefit animal welfare, Vanesa is concerned about the unintended consequences.
These range from having to manage large amounts of information, to constant tech alerts via mobile phone, and the risk that technology creates 24/7 office farming.
“There are 14 indicators on wellbeing as defined by the OECD [Organisation for Economic Co-operation and Development], and work is one of the most central features of our lives and an important indicator of our wellbeing,” she says.
Job satisfaction is measured by the demands of the job, support available and the amount of control someone has, she adds.
“With long, unsociable hours and periods of high and low intensity [in farming], plus our individual expectations of work, how we construct all of this as we grow up determines our norms and values.
“We will interview [20] farmers to understand their use and thought processes of tech: have they used it/stopped using it/are thinking of using it.
“We hope to find out why, and how, wearables are of use to farmers, and this will feed into a larger project in future.”
Six steps to data-led decisions
Data-led decisions are important for maximising accuracy and efficiency, but data need to be managed through six steps to achieve success.
Prof Robert Van Saun of Penn State University says that unless farm data are transformed into information and used, “it’s just numbers”.
The six steps are:
- Collation of data – this involves observing/measuring/recording
- Organisation – structuring the data set
- Processing – data entry and calculation produces useful information
- Reporting – presenting data for interpretation
- Integration and interpretation – putting all the data together, not looking at things in isolation
- Utilisation – the important part: using data to make good decisions for the farm.
Note: Steps 1 to 4 can all be done by herd management software. Source: Prof Robert Van Saun, Penn State University
Dr Mazdak Salavati, Dr Rob Morrison, Dr Christos Tachtatzis, Dr Vanesa Fuertes, and Prof Robert Van Saun were speaking at the recent Hannah Dairy Research Foundation Next Generation Data-Led Dairying, Edinburgh.