For systems with limited knowledge, identifying system properties from observable data is vital to decision-making. We develop data-driven tools to estimate the system transition, winning set (with respect to the control objectives) and safe operation ranges, as well as to synthesize feasible controllers.