This document summarizes research into developing new thermoelectric materials through high-throughput computing and statistical learning. Key points: - Thermoelectric materials can convert heat to electricity but require high figure of merit (ZT). Computational screening of over 50,000 compounds has identified promising candidates. - TmAgTe2 and YCuTe2 were discovered through this method and experimentally validated, with zT of 0.75 for YCuTe2. - Bournonite materials like CuPbSbS3 have low thermal conductivity but require improved electrical properties. Over 300 substitutions were modeled to explore variations. - Open data resources like Materials Project and tools like MatMiner being developed to enable data mining and accelerate materials