This dataset contains supplementary data and utilities of the publication "Coercivity influence of nanostructure in SmCo-1:7 magnets: Machine learning of high-throughput micromagnetic data" (Yang, 2026). This dataset documents homogenized anisotropic thermal conductivity of the corresponding microstructure (identified by volume fraction (Vf) and aspect ratio (Ar)) as a tensor with normalized interface thermal resistance (see Table 1). Voxelized digital microstructures and utilities are also attached for visualizing the overall thermal anisotropy. Table 1: The geometrical and thermal parameters employed in the generated microstructures. Parameters Value (Unit) Type Increment Minor principal axes length 5 (nm) Constant - Aspect ratio, \(A_\mathrm{r}\) [1 ,6] Linear 1 Inclusion volume fraction, \(V_\mathrm{f}\) [5, 60] Linear 5 Thermal conductivity ratio, \(K_\mathrm{r}\) [15, 100] Linear 15 Normalized interface resistance, \(\tilde{R}_\mathrm{s} \) [1e-6, 1e10] Logarithmic 1e2 Notice: The digital microstructure has been voxelized and stored in the ExodusII format, which can be loaded and visualized by the post-processing software, such as ParaView. In order to perform the homogenization, interface smoothening is required, i.e., to generate diffuse interfaces. In this work, we smoothened the interface by operating transient Allen-Cahn calculation with finite timesteps. Sec. 2.1 of the publication for more information).