I’m trying to make an unsupervised clustering algorithm to create a representation of chemical formula that clusters based on properties. In order to this the entire data-set would be preferable to provide as much data as possible. Is this simple to do?
Pretty straightforward, you can use
MPRester.query to request large quantities of data over essentially the entire database (minus the bandstructure, xas, and surface datasets, which have custom query methods). Please e-mail [email protected] if you’re planning on issuing heavier api queries on a regular basis.