Embodiments described herein relate to apparatuses and methods for incident prediction alerts for transmission to a health care organization system by applying rules to data sets. Each rule may define a set of data elements linked to an incident, and a processor may detect one or more sets of data elements in the data sets. The processor may normalize the data feeds, generate rules on historical data, generate prediction alerts by applying rules to near-real time data feeds, train to update rules, validate and error check rules, remove statistical noise, generate visualizations for the data feeds, and receive input and feedback data.