Getty Provenance Index User Guide

1. The Remodeled Getty Provenance Index in Arches

• Information Organization in the Remodeled Getty Provenance Index

2. Searching the Getty Provenance Index

• Introduction to Search
• Kinds of Searches
• Constructing a Simple Search
• Constructing an Advanced Search
• Additional Search Filters
• Saved Searches and Exporting Search Results

3. Interpreting Search Results

• Viewing Search Results
• Previewing Resources in Search Results

4. Viewing Resources

• Resource Reports

Information Organization in the Remodeled Getty Provenance Index

Conceptual Introduction

The Getty Provenance Index in Arches provides open access to millions of resources from dealer stock books and sales catalogs through a Getty-supported, open-source database platform. These datasets provide essential information on the ownership and market histories of artworks.

In its pre-Arches form, the Getty Provenance Index represented provenance information gathered from historic documents by replicating the tabular structure of the source material as flat-file records, meaning each entry was a single, independent row without links to other data points. In the remodeled Getty Provenance Index in Arches, those flat-file records have been transformed into a linked open data system. This means each entity is uniquely identified and connected to other relevant data using controlled vocabularies and semantic connections. This remodeling from relational to graph data transforms the implicit relationships recorded in a flat-file row into an explicit, relational web of entities that consolidates people, objects, places, and events into uniquely identified resources.

In Arches, Getty Provenance Index data is generated through events. Often, but not always, these events are related to historic transfers of ownership. These events create data that populates one of nine Resource Models used in the Provenance Index: Activity, Group, Person, Physical Object, Place, Provenance Activity, Set, Textual Work, and Visual Work. These models are based on the Linked.Art metadata application profile of the CIDOC Conceptual Reference Model (CRM) reference model.

Within a Resource Model, data is further filtered into discrete Branches that function as smaller graph models within the larger whole. For example, the record of the sale of an art object could generate a new instance or link to an existing one in the following resource models:

• A Set resource for the lot, including lot identifier branch information
• A Provenance Activity resource for the transfer of ownership, including valuation branch information
• A Physical Object resource for the art object, including material branch information
• A Person resource for the buyer, including name branch information
• A Textual Work resource for the record of the sale, including page branch information

For more information about how Arches defines Resource Models, Branches, and Nodes, see Arches Modeling Documentation.

Resource Models

Activity: Events and sub-events that can lead to transfers of ownership, such as auctions, exhibition sales, and lotteries. Activities also include the parts of those broader events that relate to the individual object sets, such as the offer of an individual lot at auction.

Group: A corporate body, such as an institution or business, or a group of individuals defined by a shared characteristic.

Person: An individual actor associated with an object, such as a creator, collector, or appraiser.

Physical Object: Works of art, material culture, or documents. These are human-produced objects that exist or existed in the physical world.

Place: Geographic location associated with a person, group, or activity.

Provenance Activity: Activities related to ownership, including transfer of ownership, transfer of custody, and inventorying of physical objects.

Set: A grouping of one or more physical objects recorded as a single lot or entry in a sale catalog or inventory.

Textual Work: Text carried by physical documents, digital resources, or parts thereof. Examples include text from sales catalogs, inventories, and stock books. A textual work can represent the text of a whole document or the text of an individual page or entry in a document.

Visual Work: The conceptual, visual content of an artwork that can be carried by one or more physical objects or digital resources. The content of an artwork remains even if the last carrier is destroyed.

Introduction to Search

Elements of Search

The “Search” function of the Getty Provenance Index in Arches allows users to query Provenance Index resources to locate data of interest. It includes a minimal design, with a Search Bar in the upper left, Search Tools in the toolbar at the top of the page, Search Results on the left, and the Search Panel in the center. Upon creating a search using the Search Bar or Search Tools, links to resources, called “cards”, populate Search Results. Search Tools can be used to further refine, visualize, or export results, often in the Search Panel.

Search Bar

The Search Bar allows users to construct Simple Searches and acts as an aggregator of all enabled filters, including those in Advanced Searches. Logographs, diacritics, and unicode special characters are also supported in search.

Search Tools

The Search Tools area provides access to the Advanced Search, Related Resources, Resource Details, Saved Search, Time Filter, and Exporting Search Results tools mentioned throughout the User Guide.

Search Results

For any search (Simple Search or Advanced Search), a filtered list of updated resources appears in Search Results. Results are displayed as cards, which list the name of the resource and a relevant subfield. Clicking on the name of the resource in a card opens that resource in a new tab. Below the resource name, two buttons open Search Tools resource previews: “Resource Details” and “Related Resources”. Clicking on “Resource Details” refreshes the Search Panel to show an overview of the resource. Clicking on “Related Resources” refreshes the Search Panel to showcase resource connections in an interactive graph. For more information on resource previews, see Previewing Resources in Search Results.

Search Panel

The content of this panel changes depending on the Search Tool selected. By default, the Search Panel directs users’ attention to the Advanced Search tool.

Kinds of Searches

Simple Search

Simple Search allows for term queries of the database using one or many keyword(s). Other filtering options allow for additional selections based on resource type. The search bar acts as an aggregator of all enabled filters. For more information, see Constructing a Simple Search.

Advanced Search

Advanced Search allows for more complex and precise queries beyond simple keyword searches. Advanced Search queries are built using data attributes known as facets, refining or expanding searches based on the structured location of information in the data model, data types, and other query conditions. For more information, see Constructing an Advanced Search.

Mixing Simple & Advanced Searches

Simple and advanced searches can be combined, allowing for general keyword search within all exposed areas of an advanced search.

Constructing a Simple Search

Keywords

Simple Search can be used to match strings of text in any branch describing resource instances. Multiple keywords can be entered in the search bar to run an “AND/OR” query. For example, a search for the two separate keywords, Caravaggio and Dream, will return the Physical Object resource Caravaggio's Dream, along with any other resources containing either Caravaggio or Dream or both.

Simple Search can also activate a “NOT” search by clicking on an entered keyword. Doing so will add a minus sign (-) before the keyword(s) and filter for all resources that do not contain the selected terms. For example, - contains term: Caravaggio will return all results that do not include the term Caravaggio in the resource.

The Simple Search Bar is complemented with search support functions to assist and refine basic keyword queries. Upon typing keyword(s), a list of term matches containing the given keyword(s) will appear as a panel below the search bar.

Click the ‘x’ next to a keyword or filter to clear it. All search query elements can be cleared by clicking “Clear Filters”.

Search Operators

Arches Simple and Advanced Search also support a number of operators to find and retrieve information.

• Exact: Search terms within double quotes (“ ”) execute an Exact prefix search. An Exact search only finds values that match the entire search string exactly, including case. For example, a Simple Search for “landscape” returns the Physical Object resource landscape and not the Physical Object resource Six landscapes…various.

• Single Character Wildcard: Search terms using question marks (?) execute a single character wildcard prefix search. For example, a Simple Search for st?ck returns the Activity resource Stock List Event D-A255 and the Physical Object resource Stuck but not the Physical Object resource Landscape – Man with Sticks, &c.

• Multi-Character Wildcard: Search terms using asterisks (*) execute a multi-character wildcard search. Using the above example, *st?ck* returns the Activity resource Stock List Event D-A255, the Physical Object resource Stuck, and the Physical Object resource Landscape – Man with Sticks, &c.

Resource Type Facet

Below the Search Bar, the “Resource Type” drop-down menu allows users to facet by Resource Model. Resource type facets can be added with keyword(s) to query for search terms within a given resource model. For example, a search for Caravaggio’s Dream can be refined to only return Physical Object results by selecting “Physical Object” under the Resource Type drop-down menu.

Constructing an Advanced Search

Facets

The Advanced Search tool features facet search for multidimensional filtering, enabling users to select multiple criteria simultaneously to execute complex queries. Advanced Search facets are generated from the Getty Provenance Index Resource Models to allow for filtering at the model, branch, sub-branch, and field level. Building an Advanced Search from one or many facets of the Provenance Index data model provides additional precision when compared to a Simple Search.

The Advanced Search interface has two sections: Search Facets on the right, which lists branches and sub-branches available for query, and the Search Panel in the center, where search criteria are added to selected facets. Adding and removing search criteria to facets in the Search Panel automatically refines resources appearing in Search Results.

To begin an Advanced Search, users must select a facet, either by navigating to the desired branch or sub-branch facet using the accordion buttons or via search at the top of the Search Facets area. Once a facet is selected, the Search Panel will show the selected branch and fields for adding search criteria. To help select relevant branch and sub-branch facets, users should consult the names and descriptions of branches, sub-branches, and fields in the Getty Provenance Index data model documentation.

Different fields allow for different kinds of search criteria according to the data model, including strings, dates, integers, concepts, and resource-instances. Strings, dates, and integers can be entered manually, while concepts and resource-instances must be selected from a typeahead list. Query conditions, including value comparisons (=, <, >, etc), Has Any/No Value, Has Any/All/No value in List, References/Does Not Reference are available. Additionally, Like/Not Like query conditions support keyword-style queries while Equals/Not supports complete phrase string queries. As with Simple Search, Search Operators can also be used in Advanced Search queries.

Multiple facets can be selected in an Advanced Search from the Search Facets panel. As additional facets are added to the Search Panel, a Boolean search operator will appear between selected facets, allowing users to define AND/OR facet relationships in their queries. Individual Advanced Search facets can either be removed by clicking the “Remove” button in the relevant facet in the Search Panel and all Advanced Search facets can be removed by clicking the “x” next to “Advanced Search” in the Search Bar or by clicking “Clear Filters” below the Search Bar.

Additional Search Filters

Time Filter

The Time Filter, indicated by a calendar icon in the Search Tools bar, allows users to refine queries by faceting on custom or predefined date ranges for resources with branches that have timespan data types. To facet on a resource branch, the “Type” dropdown menu displays all branches with timespan data types for selection. To refine results within a custom date range, users can manually define dates in ISO-8601 format or select dates in a pop-up calendar. Finally, a Time Wheel interactive allows users to see and select available timespan ranges to facet search queries on.

Saved Searches and Exporting Search Results

Saved Search

Located under the bookmark icon in the Search Tools bar, Saved Search allows users to generate preconfigured Simple and Advanced Searches. These pre-filled queries apply a variety of Simple and Advanced Search criteria that can be narrowed or expanded by users, providing useful search parameters for data exploration and refinement. Saved Search is managed by Getty Provenance Index editors and is periodically updated with new queries.

Additionally, every query, including Saved Searches, produces a stable and unique URL that can be externally saved by users.

Exporting Search Results

Search Results can be exported in a variety of formats and configurations by using the Export Search Results tool. Once users have applied search criteria to create a desired set of Search Results for export, clicking the Export Search Results button in the rightmost corner of the Search Tools bar will open a dialog for defining export format and settings. The only required criteria to create an export is the data format. Users can also define desired coordinate precision, add a report link to their export, name their export, and have their export sent to an email address. Once export settings have been set, clicking “Download” will start the export and save the files to your local machine.

Viewing Search Results

Cards & Resources

As seen in Simple Search and Advanced Search, search options are dynamic and responsive, meaning that the Search Results update synchronously as a query is built in the Search Bar. With each added search criterion, the number of relevant resources will appear above the list of cards in the Search Results area. Sort results using the drop-down menu below the Resource Type menu in three ways: Unsorted (random order), Ascending (alphabetically A-to-Z), or Descending (alphabetically, Z-to-A).

Search results appear as cards. A card provides a resource preview, including the resource name and a short description field (Note: cards will show “Undefined” when their resource description field is blank). Clicking the resource name in a card opens the corresponding resource in a new tab that contains all data instances related to the selected resource. For more about resources and the information they contain, see Resource Reports.

Troubleshooting

When a Simple or Advanced Search produces “Results: 0 of 0”, this indicates no resources match a given query. In this instance, users should make sure their queries match the contextual bounds of the Getty Provenance Index and double check selected facets, query conditions, and multi-facet Boolean operators.

Previewing Resources in Search Results

Resource Details

Resource Details, accessible by clicking “Details” on a Search Results card, allows users to preview a selected resource in the Search Panel.

Related Resources

View relationships between resources by clicking “Related Resources” on a Search Results card. Doing so activates the Related Search Tool in the Search Panel and displays an interactive network graph showcasing resource connections as defined by the semantic structure of the Getty Provenance Index data model. There are two methods for viewing Related Resource:

1. Table view: Each related resource is a row.
2. Network graph: Resources are shown as nodes in a graph, with relationships (edges), connecting them. This graph does not show geographical nor chronological connections.

In the network graph, five buttons control graph display and manipulation:

• Click a Node/Edge for Info: While selected, clicking any node or edge opens the “Info” panel, which displays additional information about the resource and its relationships; close the panel by clicking the “x” in the top-left corner
• Click a Node to Show More Relationships: Expands the graph to reveal all related nodes/edges for the selected node
• Click a Node Refocus: Recenters the graph to show only relationships connected to the newly selected node, removing any indirect relationships previously displayed
• Click a Node/Edge to Remove: Deletes the selected node or edge from the network graph
• Refresh: Recenters and resets the graph to its default state

Next to the network graph:

• “Info”: Displays details for the selected node/edge
• “Search”: Allows string searches for any currently exposed node
• “Legend”: Provides a visual explanation of node color coding used in Related Resources

Resource Reports

Conceptual Introduction

Instances of a Resource Model (known as Resources) capture information about Getty Provenance Index data and its relationships to other resources. Resources consist of discrete branches that function as smaller graph models grouping together thematically related data like names, dimensions, or timespans. Branches consist of nodes and relationships based on semantic triples (see Arches Modeling Documentation for more info). Resource Reports capture instances of Resource Model branches and relationships between resources in a series of tables.

Accessing Resource Reports

Access Resource Reports from Search by clicking on a card resource name in the Search Results. Doing so opens a corresponding Resource Report in a new tab that contains all modeled information available in the Getty Provenance Index about a given resource—such as name, type, location—and its relationship with other instances of resource models—such as Person, Physical Object, or Provenance Activity—in the Provenance Index.

Parts of a Resource Report

Resource Reports are designed with a similar layout for all resources in the Getty Provenance Index and consist of 4 sections from top to bottom:

1. Resource Header: At the top of a Resource Report, the Resource Header summarizes important resource information and displays tools for exporting resource data in a variety of formats.
2. View Selector: Two options are available: “Data” and “Related Resources”. The default view, “Data”, shows all modeled information for a resource. “Related Resources” displays other resources that are semantically linked to the current resource.
3. Section Navigation: A series of buttons specific to each Resource Model allow users to navigate to related groups of branches (sections).
4. Resource Information: Organized into branches, sub-branches, and nodes that describe a resource through its relations (edges) with other nodes (e.g., strings, concepts, timespans, resource-instances).

Branches or sub-branches are represented as tables in Resource Reports, with columns representing branch nodes and rows as individual instances of node data. For example, an artwork with two titles would create two rows of node data in the Name branch table of a Physical Object Resource Report.

Some branch tables contain sub-branch information, which can be explored in the branch table drawer by clicking on the down caret “v” on the left side of any given row. Branches without down carets do not have sub-branch data and rows with no data and a down caret only contain sub-branch data.

Within branch tables, data types include:

• Strings, numbers, dates: Encoded information from source documents (e.g., dates follow ISO-8601)
• Concepts: Controlled vocabulary terms managed by the Arches thesauri and collected in the Getty Provenance Index data model documentation
• Resource-instances: Semantically linked resources, allowing navigation through the Getty Provenance Index graph from within a single Resource Report

Resource Actions

Within a Resource Report, users can:

• Search within branch and sub-branch tables using the search bar in the top right of each table
• Sort note data alphabetically by clicking the dual arrows next to the name of each column of node data

All resources in the Getty Provenance Index have unique resource identifiers (URIs) built into their URLs, making any externally saved links stable for future reference. Additionally, users can export resources as CSV, JSON, or JSON-LD via the “Export as” options in the top right of the Resource Header. Users who find incorrect or incomplete information in the Getty Provenance Index are encouraged to submit corrections by emailing the resource URL and supporting documentation to the Getty Provenance Index Department.