Skip to content
GraphTheory.in

How we build this directory

Sources, discovery, ranking, freshness — and the limits we acknowledge.

Sources we use

Two open scholarly databases. We do not scrape Google Scholar, ResearchGate, or any platform whose terms of service prohibit automated access.

OpenAlex

CC0-licensed open index of scholarly works. We pull researcher profiles, affiliation history, citation counts, h-index, and the full publication list per researcher. We use cursor pagination at 100 results per page and identify ourselves via the polite-pool mailto parameter.

docs.openalex.org

ORCID

Researcher-curated identifiers. For researchers with an ORCID iD attached to their OpenAlex profile, we enrich the record with biography, employment history, education, awards, memberships, and works metadata from their public ORCID record.

info.orcid.org

How we discover researchers

How a new researcher enters the directory.

  1. 01

    Keywords map to OpenAlex concepts

    Editor-managed keywords (e.g. "Graph theory", "Combinatorics", "Network science") are resolved to OpenAlex concept IDs. Unmapped keywords are resolved on the fly during the next harvest run.

  2. 02

    Concepts + India filter discovery

    OpenAlex's group_by query returns authors who have published works tagged with our concepts and affiliated with at least one institution whose country code is IN. We dedupe against the existing roster.

  3. 03

    Per-researcher enrichment

    For each new author we fetch their profile, paginate through their works (filtered to the same concepts + India), then look up their ORCID public record. Each step respects daily request budgets and resumes automatically the next day if a budget runs out mid-flight.

How we rank

The default /researchers page uses an editor-defined ordering: manually pinned researchers first, then everyone else ordered by h-index. Visitors can sort by h-index, total citations, or publication count via the sort controls. The metric chosen is shown in the URL so links are shareable.

Sort options

  • Editor's order— pinned first, then h-index. The default landing-page order.
  • h-index — the largest n such that the author has at least n papers each cited at least n times.
  • Citations— total citation count across all indexed publications.
  • Publications— total publication count (filtered to the indexed concepts + India).

We don't publish a composite score. The metrics are presented separately so you can pick the lens you care about.

How fresh the data is

Predictable cadence, polite to upstream APIs, and resilient to rate limits.

  • Monthly cadence

    A combined harvest runs once a month: discover new researchers, refresh existing rows, and catch up on any pending ORCID enrichments. Admins can also trigger Discover or Refresh on demand for any selection of keywords.

  • Daily request budgets

    OpenAlex and ORCID each have a daily request cap (with a configurable safety margin). The harvester checks the cap before every request and reconciles against upstream rate-limit headers when available, so we never accidentally overrun a quota.

  • Defer + auto-resume

    If a budget is exhausted mid-run, the harvester pauses cleanly at the current researcher boundary, persists state, and a 15-minute cron auto-resumes once the daily budget resets at UTC midnight. No state is lost; nothing is half-saved.

What we don't claim

The portal is built on open data and is explicit about its scope. We do not verify employment claims, recompute citation metrics, peer-review publications, or score researchers ourselves.

  • No OpenAlex profile, no entry. Researchers without an OpenAlex author record are not indexed. If you publish in venues OpenAlex doesn't cover, you won't appear.
  • Author disambiguation is imperfect. OpenAlex's automated algorithm is approximately 90–95% accurate. For common name combinations, errors do occur — papers can be mis-attributed.
  • ORCID coverage is uneven. Researchers who haven't connected an ORCID iD won't have biographical enrichment, even if they're fully indexed in OpenAlex.
  • Publications without DOIs may be approximate. Citation counts and venue metadata are most reliable for papers with DOIs. For older or non-DOI works, treat the numbers as best-effort.

How to flag a problem

The portal gets better when researchers and readers tell us what's missing, wrong, or out of date.

  • Claim your profile

    If you're listed or want to submit your profile as a new researcher, you can claim it here to set up your dashboard.

    Claim/Submit profile →

  • Report a correction

    Spotted incorrect data, wrong affiliation, or a misattributed paper? Get in touch and we'll fix it.

    Contact us →