Digital data sharing leads to unprecedented opportunities to develop data-driven systems for supporting economic activities, the social and political life, and science. Many open-access datasets are RDF graphs, but others are CSV files, Neo4J property graphs, JSON or XML documents, etc. Potential users need to understand a dataset in order to decide if it is useful for their goal. While some datasets come with a schema and/or documentation, this is not always the case. Data summarization or schema inference tools have been proposed, specializing in XML, or JSON, or the RDF data models. In this work, we present a dataset abstraction approach, which () applies on relational, CSV, XML, JSON, RDF or Property Graph data; () computes an abstraction meant for humans (as opposed to a schema meant for a parser); () integrates Information Extraction data profiling, to also classify dataset content among a set of categories of interest to the user. Our abstractions are conceptually close to an Entity-Relationship diagram, if one allows nested and possibly heterogeneous structure within entities.