In high-efficiency power plants, some components must withstand high stress and temperature. Materials scientists manipulate alloying element composition and thermomechanical processing to design specific mechanical properties in alloys. The analyzed 8-12% Cr steel dataset, for iron base alloy compositions (>80) and processing parameters, displayed results of tensile strength in 34 columns by 915 rows. To address non-linearity of the tensile properties, data analyses were carried out in composition-based clusters. We illustrated the effectiveness of the clustering approach in both classification and non-linear regression problems. This data science approach can assist with domain-guided statistical design for optimal manufacturing, computational materials engineering, uncertainty quantification to support decision-making, and additional scientific insight into complex, noisy, high-dimensional, and high-volume data sets. The hypotheses generated via the non-linear data analyses were tested on extended compositional ranges, providing new and interesting insights. These insights produced by data science can be interpreted using domain science knowledge for further validation and knowledge discovery.