SPEEDD (Scalable ProactivE Event-Driven Decision making) will develop a system for proactive event-based decision-making: decisions will be triggered by forecasting events -whether they correspond to problems or opportunities- instead of reacting to them once they happen. The decisions and actions will be real-time, in the sense that they will be taken under tight time constraints, and require on-the-fly processing of "Big Data", ie extremely large amounts of noisy data storming from different geographical locations as well as historical data. The effectiveness of the SPEEDD solution will be evaluated against concrete requirements in traffic management and credit card fraud detection.SPEEDD will contribute to the state of the art by:-Developing techniques for on-the-fly, low-latency, scalable, distributed monitoring and online distributed learning, given extremely large, geographically distributed, noisy event streams, and highly complex event patterns.-Developing novel methods for real-time event recognition and forecasting that are resilient to various types of uncertainty and supported by novel machine learning techniques for continuously improving recognition and forecasting accuracy.-Providing novel methods for event-based, real-time decision-making under uncertainty.-Developing techniques for real-time explanation and visualisation of Big Data.The expected outcome of SPEEDD includes:-Technology for real-time event recognition and forecasting under uncertainty.-Technology for event-based, real-time decision-making under uncertainty.-Visual analytics suite for real-time interaction with, and explanation of, Big Data, as well as proactive decision-making support.-A highly scalable proactive event-driven computing prototype integrating all SPEEDD components. The prototype will support a Human Factors validation by means of usability and effectiveness testing.-A suite of real-world demonstrations using live data in an operational environment.