Source: The University of British Columbia
Data is at the heart of all research, and there is no shortage of data points at UBC’s Dairy Education and Research Centre (DERC). The researchers at DERC collect a wide range of data from their herd. Each cow wears one or more devices that measure(s) everything from how much they eat and rest, to where they spend their time, and even who else in the herd they hang out with!
Visualizing Years of Data
Throughout the years, the DERC has built an impressive database of information. However, combing through the data is time consuming, and access to the time and skills of a data scientist is often limited. Presented with this issue, the DERC researchers have worked with the Masters of Data Science (MDS) program to create a new tool to better understand the lives and welfare of their dairy cattle.
Introducing, Peek-a-Moo, a data visualization tool that creates a customizable and interactive dashboard to monitor animal welfare among the dairy cows.
This new tool was born from a desire to fully utilize the years of collected data at the dairy and genuine interest from UBC MDS students, who volunteered their time to develop the tool through a three-week hackathon.
Working closely with Dr Borbala Foris, a postdoctoral fellow in the Faculty of Land and Food Systems’s Animal Welfare Program, and Sky Sheng, a graduate student in Applied Animal Biology, the seven-student team of data scientists built a dashboard that allowed at-a-glance summaries of cow behaviour.
Peek-a-Moo dashboards can show:
- Timelines with every single cow’s behaviour throughout the day along with a “herd average” to detect any abnormal activity
- Social networks that display relationships between cows, positive and negative associations as well as the frequency of initiating or receiving competitive interactions
- Feed bin graphs that measure dynamic feed distributions across all bins throughout the day
- Warning messages for bin malfunction and cows with abnormal level of feed or water intake
See the Peek-a-Moo dashboard in action here:
All of this information can be toggled to select any date range or any selection of cows. Peek-a-Moo allows researchers to monitor the herd without having to watch them in person or via video. The new tool can even be used as part of an alert system, detecting early signs of illness or improvement in the cattle. Its visual nature also means people outside of the industry can easily understand findings and results from the herd.
This example of interdisciplinary collaboration will allow UBC researchers to more efficiently analyze herd behavior. With Peek-a-Moo taking care of data crunching and visualization, the researchers at the DERC have more time to focus on improving the lives of dairy cattle and the people who care for them.