Data formats

Each collection is written in one pass into several files. They hold the same records with the same ids, so you can start with the simplest format and move to a richer one later without re-keying anything.

To try the formats on real data, open the Crown Default download block, or start with the small maritime collection.

Choose a format

FollowTheMoney JSONentities.ftm.json and targets.nested.json: every property, typed, one entity per line. For loading into a database, search index or graph.Recommended Simplified CSVtargets.simple.csv: one row per record in sixteen columns. For spreadsheets and quick lookups. Plain-text namesnames.txt: every distinct name and alias, one per line. For text search tools and first-pass filters. Senzingsenzing.json: records in the Senzing entity specification, ready to load into Senzing entity resolution.

Statements CSV

statements.csv breaks every record down into single values: one row per property value, with the record it belongs to, the register it came from and when it was first and last seen. It is the most detailed file, and the one to use for audits, value-level change analysis or rebuilding entities your own way. See the statement data model.

Column	Contents
`id`	Statement id: SHA-1 of record id, property and value. Stable while the value stays the same.
`entity_id`	Record id (`crown-…`) or Sanction id (`crown-…-sanction`)
`canonical_id`	Merged id (`crown-e-…`) when the record has been matched with others; otherwise the record id
`prop`	Property name, e.g. `name`, `alias`, `birthDate`, `imoNumber`, `program`
`prop_type`	Value type: `name`, `date`, `country`, `identifier`, `url`, `text`, `topic`, `entity` or `string`
`schema`	Entity type of the record or `Sanction`
`value`	The value
`dataset`	Source register, short identifier
`lang`, `original_value`	Reserved; empty today
`first_seen`, `last_seen`	When the record was first collected and last confirmed (ISO 8601)

id,entity_id,canonical_id,prop,prop_type,schema,value,dataset,lang,original_value,first_seen,last_seen
29645e46c9b65ad8dbc0a4af828d56faaa71892d,crown-380054154,crown-380054154,name,name,Vessel,PELIKAN,paris_mou_banned,,,2026-09-24T11:38:49.104591+00:00,2026-09-24T11:38:49.104591+00:00
90c07d9bcb169c5c8c5f61562e162367cecc0992,crown-380054154,crown-380054154,imoNumber,identifier,Vessel,IMO6922169,paris_mou_banned,,,2026-09-24T11:38:49.104591+00:00,2026-09-24T11:38:49.104591+00:00

Index and metadata files

Alongside the data, each collection has small files that describe it:

File	Contents
`index.json`	Version, build time, record and register counts, entity types, countries, how many records fill each field, and every file with its size, row count and SHA-256 checksum
`sources.json`	Every register in the collection with its record count, country, publisher type, refresh interval and source link
`statistics.json`	The figures of every build, newest first (shown on the statistics page)
`delta.json`, `deltas.json`	Changes since the previous build, and the list of recent builds with their delta files (see Delta files)

Two files cover all collections: the catalogue /downloads/collections/index.json, and the data model /downloads/meta/model.json, which lists the schemas, properties, topics and file layouts used in the exports.

Files in each collection

Collection

File	Type	Rows	Size
Loading…

Keeping up to date

Files are replaced in place after each build. Keeping data fresh explains how to tell when a new build is out and how to check a download; Delta files explains how to apply only the changes.

Specialised exports

Maritime. The maritime collection holds vessels and every record carrying an IMO number, in all the formats above. See Maritime and vessels.
Senzing. Every collection has a senzing.json; see Senzing.
Securities. We do not publish a securities export yet; see Securities.

Need a different layout or a subset? See Custom exports or talk to us.

Last updated 2026-10-03