If the timing of an event can be predicted in advance, it becomes possible to mitigate its risk or fully leverage its opportunity through appropriate proactive actions. For example, preventive maintenance can be performed before equipment failure, or discount coupons can be distributed just before a customer’s visit to encourage purchases. The framework for analyzing event-related data and predicting their occurrence timing is known as event series analysis. In this talk, I present recent advances in identifying the underlying mechanisms of event occurrence and forecasting future events through machine learning-based approaches to event series analysis.