Presentation Information
[8p-E203-8]Starrydata Auto-Explorer: An Overview Tool for Thermoelectric Literature Data
〇Yukari Katsura1,2,3, Eiji Koyama1, Dewi Yana1, Sundas Rani1,2, Atsumi Tanaka1, Tomoya Mato1, Masahiko Demura1 (1.NIMS, 2.Univ. of Tsukuba, 3.RIKEN)
Keywords:
thermoelectric materials,informatics,principal component analysis
We developed Starrydata Auto-Explorer, a program that applies principal component analysis (PCA)-based visualization of chemical compositions, in a single batch operation, to material systems specified by the user, drawing on a dataset of thermoelectric material entries. When the user specifies the groups of elements occupying each site of the parent material along with their composition ranges, the program extracts the corresponding set of samples from Starrydata's tens of thousands of data entries and generates a composition PCA map, k-means clustering, and a property scatter plot for each sample, all as a single dashboard image.
