RELEASE_NOTES.txt

OPENALEX STANDARD-FORMAT SNAPSHOT RELEASE NOTES

RELEASE 2026-02-03
- additional data quality improvements in multiple entities

RELEASE 2026-01-15
- multiple data quality improvements in works and authors
- fix author bug where deprecated authors were assigned to works

RELEASE 2025-11-12
- switch to walden dataset which is now available at s3://openalex/data
- works, authors, institutions, etc are from walden which is the default in the API
- includes "xpack" records in works, so full works count is 463M

RELEASE 2025-09-30
- added new works

RELEASE 2025-08-21
- added new works

RELEASE 2025-07-07
- added new works

RELEASE 2025-05-30
- added new works

RELEASE 2025-05-07
- added new works

RELEASE 2025-03-31
- added new works

RELEASE 2025-02-27
- added new works

RELEASE 2025-01-29
- removed 350k abstracts with invalid or junk content
- after this release, the snapshot will be updated quarterly

RELEASE 2024-12-31
- used new ROR matching algorithm to assign affiliations to institutions with zero works; 7.6k additional institutions now have work affiliations
- added affiliations to 4.5M authorships

RELEASE 2024-11-25
- added 3.5 million new author IDs after fixing bug introduced by ORCID integration
- added 103 new sources (journals) that started publishing in latter part of 2024
- ingested 145 author change requests with new curation form
- added/removed institution affiliations for 3.5k works based on works-magnet curation requests
- processed 50 source curation requests
- curated 5 institutions that were lacking affiliations

RELEASE 2024-10-31
- detect additional paratext
- fixed author alternate names bug

RELEASE 2024-09-27
- added ~14M affiliations to works
- adjusted year retrieved from Crossref, using the earliest from issued, published, approved, created, deposited. This affects ~20M works.

RELEASE 2024-08-29
- add new object citation_normalized_percentile to works, which is a percentile rank of citations normalized by the number of works in the same year and subfield
- add more references to works using previous MAG snapshot
- restored some missing affiliations

RELEASE 2024-07-30
- new work type: retraction
- used data from Pubmed to reclassify 4M works from type "article" to one of: editorial, erratum, letter, preprint, review, retraction
- improved type classification for works using string matching
- change titles (display_name) for ~30k journals based on data from Crossref
- delete 187,452 works: deleted Zenodo records. (merge into deleted id: W4285719527)
- clean up author names: remove non-name strings prepended to certain author names; remove non-printing characters and whitespace, delete authors with bad names (only whitespace, only numbers)

RELEASE 2024-06-30
- add new affiliations field to Work.authorships, allowing more detailed mapping of raw affiliation strings to institutions
- new is_core boolean added to sources and associated works based on dataset from CWTS: https://zenodo.org/records/10949671
- fixed bug causing some works' OA status to be out of sync with Unpaywall
- APC estimates (apc_paid) are no longer given for works with OA status of closed, bronze, or green

RELEASE 2024-05-30
- added 151M new references to works (7.61% increase) by matching references without DOIs using title/author/publication year
- updated authorship information for 17.9M works by syncing Crossref changes
- added 4 new work types, reclassifying existing works: “preprint” (5.7M), “libguides” (1.8M), “review” (820k), and “supplementary-materials” (50k)
- "super system" institutions such as University of California System are removed from institution lineage
- added datasets from DataCite: 1.07M from Cambridge Structural Database and 709k from Harvard Dataverse

RELEASE 2024-04-25
- added affiliations to ~2.5M works using open access PDFs parsed by grobid
- ingested 3.4M works from DataCite, primarily from Zenodo, Arxiv, and Figshare
- fixed language detection bug that occurred when title and abstract all uppercase
- override language assignment for some major English-language journals
- remove Author.last_known_institution in favor of Author.last_known_institution (in progress)
- remove Work.authorships.raw_affiliation_string in favor of Work.authorships.raw_affiliation_strings (in progress)

RELEASE 2024-03-27
- set up automatic updates for Retraction Watch (see https://doi.org/10.13003/c23rw1d9 for info about Retraction Watch)
- marked ~2k additional works as retracted, and corrected 1-2k works that were incorrectly labeled is_retracted
- add siblings to domains, field, subfields

RELEASE 2024-02-27
- added topics to works
- modified ID format for topic domain, field, and subfield within works, from integer to openalex string. About 60% of
    works have the new format. The remaining 40% will be updated by the next release.
- added new topics, domains, fields, and subfields entities to the snapshot
- fixed host_organization bug within sources

RELEASE 2024-01-24
- added Author.last_known_institutions, a list of institutions for the affiliations of the author's most recent work (last_known_affiliation will be deprecated in the future)
- added indexed_in to works
- merged around 300 institutions based on ROR data
- removed the license type "publisher-specific, author manuscript" from ~140k work locations, changing them to either "publisher-specific-oa" or closed

RELEASE 2023-12-20
- added Author.affiliations, an author's 10 most recent associated institutions and years of publications
- merged 566k duplicated works associated with HAL repository
- removed old author IDs (ID less than 5000000000) from merged_ids/authors
- changed cited_by_percentile_year in works to an integer

RELEASE 2023-11-21
- improved affiliation matching for over 2 million works
- added keywords to works
- added cited_by_percentile_year to works
- removed 3.9 million authors with 0 works (merged into deleted profile)
- improved oa status classification, converting 2.1 million closed works to open access statuses
 (ongoing - credit: https://subugoe.github.io/scholcomm_analytics/posts/oalex_oa_status/)

RELEASE 2023-10-18
- added new works
- added more sources

RELEASE 2023-09-20
- added raw author name to authorships objects in works
- institution lineage (parent institution IDs) available in works, authors, institutions
- sustainable development goals assigned to 209 million works
- improved institution matching for 1.1 million works
- countries distinct count available in works
- added ~700 new sources
- matched primary source for 248,647 old works
- abstract inverted index is correct object in snapshot (InvertedIndex key removed)
- updated_dates are in full ISO format
- documentation scripts updated for current snapshot

RELEASE 2023-08-18
- released new authors disambiguation feature
- fixed missing source assignment for 5.7M works
- improved affiliation matching resulting in additional ~1.1M works matched to institutions
- works with more than 100 authors no longer have authors truncated
- modified Work.type, added Work.type_crossref
- added APC data for 3,508 journals
- added authorships.countries attribute
- resolved minor snapshot bugs affecting abstract_inverted_index and manifest, removed "@" fields

RELEASE 2023-07-11
- add references_count to works
- add records across all entities
- updated and improved records across all entities

RELEASE 2023-06-02
- add new works
- add apc_payment to works
- add locations_count to works
- improved coverage of alternate titles, homepage, country code for funders, publishers, and sources

RELEASE 2023-05-03
- add funders entity
- add grants to works
- add APC payments to sources

RELEASE 2023-03-28
- truncate work display_names to 500 characters
- truncate author display_names to 100 characters
- added summary stats for every entity type except works
- added new publishers and works

RELEASE 2023-02-21
- merged ~170 million authors into deleted author record as part of disambigation project
- renamed venues to sources
- add publishers entity
- added new works

RELEASE 2022-12-21
- the values in host_venue have been added to alternate_host_venues in works, which paves way for the new locations
list that will contain all possible venues for a work
- added new works

RELEASE 2022-11-14
- new fields in venues: type, apc_usd, alternate_titles, abbreviated_title, fatcat_id, and wikidata_id
- added new works

RELEASE 2022-10-10
- implemented automated concept tagger v3, which provides complete paths to concept level 0
- restored 170M "lost" citations that were in MAG and we deleted
- added new works

RELEASE 2022-09-16
- added 1.3 million new Works
- removed 700 thousand duplicate Authors

RELEASE 2022-08-09
- removed duplicate institutions for each author in Work.authorships
- made DOIs unique across Works. removed incorrect DOIs from 500K works and merged 1M sets of works with the same DOI.
- added 900K new Works

RELEASE 2022-07-09
- added new papers and corresponding data
- added missing related works
- updated many concepts using improved algorithm
- updated many affiliation mappings using improved algorithm
- removed duplicate Authors and Works

RELEASE 2022-06-09
- added new papers and corresponding data
- added about 28 million papers with Crossref DOIs
- added 23 thousand new journal webpage links, thanks in large part to DOAJ data
- fixed a bug with cited_by_year in venues
- new works now use an improved algorithm for mapping affiliation data to ROR IDs
- merged some duplicate institutions and venues (all references have been consolidated to one of the IDs)

RELEASE 2022-05-12
- added new papers and corresponding data

RELEASE 2022-04-30
- implemented automated concept tagger v2, which uses more fields to assign concepts to works
- added new papers and corresponding data
- updated 8715 venues that had issn_l listed but no issns in "issns" key

RELEASE 2022-04-07
- added new papers and corresponding data
- add DOIs to 1,242,303 existing works
- new related works to many works missing them

RELEASE 2022-03-11
- added new papers and corresponding data
- added 1181 new journals to Venues
- updated publisher, title, ISSNs on a few hundred journals

RELEASE 2022-03-01
- added new papers and corresponding data
- added 45 new journals to Venues
- added ancestors to 1500 Concepts without ancestors
- fixed a bug with some tabs in the publisher field in Venues and Works

RELEASE 2022-02-22
- added new papers and corresponding data
- added "created_date" to all entities

Partial release on 2022-02-04 (updated Institutions and Venues)
- ensured each institution has a distinct ROR (identified some institutions that will be merged in a future release, details TBD)
- updated institution names and data to match what is in ROR
- added all ROR institutions to Institutions (about 81,000 new institutions)
- matched papers to new institutions (some errors, but will improve over time)
- updated last known institution for millions of authors
- don't show citation counts for future years in "counts_by_year"
- ensured each journal has a distinct ISSN-L (identified some journals that will be merged in a future release, details TBD)
- add many new journals to venues table, link to works when possible (about 73,000 new journals)
- add more links from works to venues using Crossref data

RELEASE 2022-01-31
- added new papers and corresponding data
- remove blank lines
- citation counts for concepts use improved algorithm

RELEASE 2022-01-24
- added work.abstract_inverted_index
- added work.affiliations.raw_affiliation_string
- changed the type of work.cited_by_api_url: was a list by mistake, now a string
- removed ids that have a NULL value from the "ids" dict for all five entity types
- corrected the spelling of institution.associated_institutions
- does not include new entities since last release: a new snapshot will be released soon with recently-published works

RELEASE 2022-01-02
Released on Jan 2, 2022 at s3://openalex/data/
- First release