Many applications today are confronted with the task of discovering useful patterns in a flood of high-speed, real-time, and continuous stream of data. For instance, monitoring of Internet traffic for usage patterns or anomalous behavior, analysis of financial data or web query logs, trend discovery in retail or credit card transactions, location-aware services, and data processing in sensor networks fit this description. In response to these emerging applications, a new model of algorithms, called \emph{data stream algorithms}, has received considerable attention recently within theory, database, and networking communities. In this talk, I will discuss some of our recent work on finding spatial patterns in multi-dimensional data streams.