Lantmännen Feed (formerly Raisioagro) has been making feed mixes since 1948. 90 % of farm feeds are delivered in bulk form directly from the production plants in Kouvola and Ylivieska to livestock producers such as dairy farmers. On dairy farms, cows carry a tag with an ID chip around their neck. In loose-house barns the cows can walk through a milking robot when they want to be milked. The robot milks the cow, and tests the raw milk for various qualities.
Lantmännen Feed wanted to utilize the milking robot data which they have been collecting from 200 farms over 3 years to understand the lactation patterns of the cows using advanced analytics techniques like machine learning. Bilot created a solution which enables farmers to optimize feeding plans for the cows, thereby improving their health by preventing overfeeding or underfeeding, which in turn results in an increase in milk yield.
We also developed an alert system that helps farmers utilize their limited time more wisely by focusing their caretaker efforts on the cows displaying unusual behavior. We built predictive models that understand the milk production patterns of the cows and found practical ways to utilize these models to achieve better outcomes.
What we did
We helped the customer with:
Data manipulation and analysis
Quick hypothesis testing
Model refinement and web service deployment
Construction and service of a self-service portal website
SQL Server Management Studio
Python with scientific libraries
SAP Predictive Analytics
Microsoft Azure Machine Learning
PHP, HTML, CSS, JS