Drug-Drug Interactions (DDIs) are a major cause of preventable adverse drug reactions and a huge burden on public health and the healthcare system. On the other hand, there is a large amount of drug-related (open) data published on the Web, describing various properties of drugs and their relationships to other drugs, genes, diseases, and related concepts and entities. In this demonstration, we describe an end-to-end system we have designed to take in various Web data sources as input and provide as output a prediction of DDIs along with an explanation of why two drugs may interact. The system first creates a knowledge graph out of input data sources through large-scale semantic integration, and then performs link prediction among drug entities in the graph through large-scale similarity analysis and machine learning. The link prediction is performed using a logistic regression model over several similarity matrices built using different drug similarity measures. We present both the efficient link prediction framework implemented in Apache Spark, and our APIs and Web interface for predicting DDIs and exploring their potential causes and nature.