Sample Provenance Datasets

Knoedler Stock Books

About the Dataset

This dataset contains 53,958 records transcribed from the 11 painting stock books and enhanced with information from the 21 paintings and watercolors sales books of M. Knoedler & Co. in New York (1872-1970).

The Getty Provenance Index

Knoedler Dataset

Knoedler Full Dataset

Documentation of Variables

(click to expand each section)

Variable Type Description Notes CIDOC-CRM / Linked Art Mapping
transaction_id String Unique identifier for the transaction event. Getty-derived identifier (URI suffix only). E7 Activity / la:ProvenanceEntry
transaction_activity_label String Human-readable label describing the transaction activity. Verbatim label summarizing the transaction. E7 Activity → rdfs:label
acquisition_id String Identifier of the acquisition sub-event within the transaction. Getty-derived identifier; represents method/context of acquisition. E8 Acquisition
acquisition_label String Label describing the acquisition type or method. Compound label (e.g. provenance + purchase). E8 Acquisition → P2 has type
transaction_date Date / String Date associated with the transaction. Recorded transaction date; may include uncertainty (e.g. day = 00). E7 Activity → P4 has time-span
artwork_id String Identifier of the artwork involved in the transaction. Getty-derived object identifier. E22 Human-Made Object
artwork_label String Title or short descriptive name of the artwork. Verbatim title as recorded. E22 Human-Made Object → rdfs:label
artwork_type String High-level type of the artwork. Typically Getty/AAT-derived textual label. E22 Human-Made Object → P2 has type
visual_object_type_labels String More specific representational or genre classification. Textual category (e.g. landscapes). E55 Type
artist_id String Identifier of the artist associated with the artwork. Getty-derived authority identifier. E21 Person
artist_label String Name of the artist associated with the artwork. Verbatim name; not normalized beyond source. E21 Person → rdfs:label
artists_nationality String Nationality of the artist, when recorded. Textual value; not reconciled to authority files. E21 Person → P107i is current or former member of
buyer_id String Identifier of the buyer involved in the transaction. Getty-derived actor identifier. E39 Actor
buyer_label String Name of the buyer involved in the transaction. Verbatim name only; locations removed. E39 Actor → P22 transferred title to
seller_id String Identifier of the seller involved in the transaction. Getty-derived actor identifier. E39 Actor
seller_label String Name of the seller involved in the transaction. Verbatim name only; locations removed. E39 Actor → P23 transferred title from
payment_quantity Numeric Monetary amount recorded for the transaction. Numeric value without currency symbol. E97 Monetary Amount
payment_currency String Currency in which the payment was made. Textual currency label (not ISO-normalized). E98 Currency

Art Institute of Chicago

About the Dataset

This dataset compiles provenance records from the Art Institute of Chicago’s online collection database.
It includes artworks such as paintings, sculptures, prints, drawings, and decorative arts.

In the original API, provenance information appears as free text. For the purposes of this workshop, this material has been transformed into structured provenance events, where each event includes one or more involved parties (people, families, institutions) and other semantic components. The information of all provenance statements remains exactly as provided by the Art Institute of Chicago — no corrections, additions, or interpretative changes have been introduced.

API — Art Institute of Chicago

Sunday Practice

Art Institute of Chicago Provenance Dataset

Documentation of Variables

(click to expand each section)

These fields sit at the top level of each JSON object.

Variable Type Description
id Integer Internal numeric identifier for each artwork (unique).
title String Public title of the work.
alt_titles String or array, nullable Alternative titles when available.
main_reference_number String Accession/inventory number.
date_start Integer Earliest possible creation year (BCE negative, CE positive).
date_end Integer Latest possible creation year.
artist_display String Display text describing maker/culture/origin.
place_of_origin String Normalized place name extracted from the record.
dimensions String Free-text measurements.
medium_display String Materials and techniques.
inscriptions String, nullable Description or transcription of inscriptions.
credit_line String How the work entered the collection (gift, purchase, etc.).
provenance_text String Narrative provenance as written by AIC curators.
events_structured Array Parsed provenance events derived from the narrative.

Each entry in events_structured represents a structured provenance fragment.

Variable Type Description
event String Event identifier (e.g., “event_1”). Groups components belonging to the same ownership step.
value String Extracted fragment referring to a person, method, date, place, etc.
type Array of strings Semantic labels defining the nature of value. May include:
• Person — Named individual
• Party — Any owner (person, family, institution)
• Group — Collective/institutional actor
• Receiver — Party acquiring the work
• Method — How the object changed hands (“sold”, “given”)
• Time — Specific dates or years
• Vagueness — Uncertain expressions (“by 1938”)
• Name — Explicit name tag (overlaps with Person/Party)

© 2025 Coding Provenance Workshop — Licensed under a Creative Commons Attribution 4.0 International License.
Reuse is encouraged — we happily promote changes of custody. Just be sure to write a good provenance record and cite this workshop.