Matryoshka NYC data lake (NYC Open Data, Blend snapshot)

The NYC data lake used in the PVLDB 2026 paper Matryoshka: 1,069 tables from NYC Open Data (data.cityofnewyork.us), in the snapshot used by the Blend data discovery system (Esmailoghli et al.). The tables are distributed as the three archives in which we received them (education, finance, services), each containing gzip-compressed TSV files named by their NYC Open Data four-by-four dataset identifier (e.g. tm6d-hbzd). Source and modifications. All tables originate from NYC Open Data and are redistributed unmodified from the Blend snapshot. NYC public data sets are made available "without any registration requirement, license requirement or restrictions on their use" (NYC Administrative Code § 23-502(d)); the source is NYC Open Data, the version is the snapshot date of the archives, and no modifications were made. The City of New York provides the data as is, without warranty. Files: datasets.data-cityofnewyork-us.{education,finance,services}.tar.gz (201, 176 and 692 tables).