yes, if the headers for the data is known and the value in the dataset revolves around the relationships between entities rather than the identifiable data. As

yes, if the headers for the data is known and the value in the dataset revolves around the relationships between entities rather than the identifiable data. As an example - an anonymized sample from facebook where emails and names are anonymized would still be useful for all other kinds of stats regarding friend groups, interest clusters etc etc