This document discusses using spatial statistics and machine learning classifiers to detect botnet command and control (C2) servers through DNS lookups. It aims to accurately detect botnet traffic with no prior knowledge, being lightweight, fast, and adaptable. The document examines using spatial measures like nearest neighbors analysis and Moran's and Geary's indices on the locations of IP addresses for DNS lookups. This is used to train classifiers to distinguish between benign and fast-flux botnet domains. The classifiers achieved over 95% accuracy and were shown to have minimal performance impact when processing 20,000 domains in under 13 seconds.