Disease and production disturbances in a herd are often silent, and may cause major losses if they remain undetected. Continuous analysis and monitoring of data already recorded at the farm level can improve profitability. This project will develop methods for monitoring trends in production and animal health indicators, and detect early signs of disease or other disturbances. The aim is to produce information that can help individual farms to avoid losses. The study will compare different methods to develop a continuous and fully automated system of analyses of production data already recorded on the farm (e.g. data on litter size/piglet mortality or feed consumption). The result is an on-farm, data-driven alarm system for disease outbreaks or other hazards. The system can also be used to monitor the farm's productivity and health. The results have the potential to be used by all producers who record similar data, and can also be adapted to other species in animal production.