Tuberculosis (TB) infects one third of the world's population and is the second leading cause of death from a single infectious agent worldwide. The emergence of drug resistant TB remains a constant threat. We examine how machine learning methods can help control tuberculosis. DNA fingerprints of Mycobacterium tuberculosis complex bacteria (Mtb) are routinely gathered from TB patient isolates for every tuberculosis patient in the United States to support TB tracking and control efforts. We develop learning models to predict the genetic lineages of Mtb based on DNA fingerprints. Mining of tuberculosis patient surveillance data with respect to these genetic lineages helps discover outbreaks, improve TB control, and reveal Mtb phenotype differences. We discuss learning- and visualization-based tools to support public health efforts towards TB control in development for the New York City Health Department.