openscience.works
Turning data into impact narratives.
30+ data sources
4 story types
407 impact stories
100% open infrastructure
Turning data into impact narratives.
openscience.works builds open, reusable services that help researchers, libraries, and publishers understand how open scholarly works are used, shared, and discussed across the research ecosystem. Drawing on open data and transparent methods, the platform turns diverse signals into clear, contextual impact narratives — not opaque scores.
Research impact has always been fragmented: usage, citations, reviews, repositories, library holdings, and online discussion are spread across dozens of disconnected platforms. Instead of asking users to chase these signals themselves, openscience.works connects them into a single, readable story for every work — from journal articles and monographs to datasets and software.
Built on transparent sources, open infrastructure, and reproducible methods, every data point in every story is traceable back to its origin.
Unified intelligence across books, articles, datasets, and software.
Connecting journal articles to policy, public discussion, and broader research visibility.
Tracking monographs, chapters, and edited volumes across usage, citations, and teaching.
Making datasets discoverable, reusable, and connected to the scholarly record.
Tracking scientific software and code adoption across the research ecosystem.
Aggregating, normalising, and contextualising research signals across the open ecosystem.
Ingest metadata and signals from 30+ trusted open data sources and repositories.
Clean and harmonise raw signals — resolving identifiers, contributors, affiliations, and duplicates.
Move beyond raw counts to understand how a work is used, taught, discussed, and cited.
Produce readable impact stories, portfolio dashboards, and API-accessible outputs.
Traditional academic metrics only tell a fraction of the story. openscience.works aggregates signals from 30+ trusted open sources across seven dimensions — APIs where available, curated feeds and citation links elsewhere.
Our pipeline goes beyond citation counts. By analysing the composition of a work's attention, usage, and citing landscape, we dynamically assign Inferred Roles across four impact domains. These are heuristic signals — not ground truth — and are always displayed with confidence context.
Roles are now normalised through a canonical role map, so legacy and current role IDs resolve to one stable label in stories, filters, and dashboards. Scoring remains type-aware: thresholds differ for articles, books, datasets, and software to reflect different signal densities.
For cross-type interpretation, compare evidence patterns and within-type salience rather than raw scores alone.
Foundational building block cited heavily in core academic journals and monographs.
Accumulated strong citations or usage signals in an unusually short period — indicating immediate relevance in the field.
Frequently cited in literature reviews and meta-analyses as a shorthand reference for a specific finding or framework.
A standard protocol or tool used by other researchers, inferred from reproducibility-oriented uptake patterns.
Explicitly applied as hard data or methodology, often accompanied by stewardship, reuse, or supporting citation context.
Integrated into durable open-knowledge surfaces and scholarly discussion layers, indicating broad interpretive uptake.
Value through direct consumption: high downloads, HTML views, or library holdings rather than formal citation.
Teaching adoption signal spanning OCW, syllabi/OER, open textbook matches, and educational video usage across story types.
Sustained engagement across open-web channels, calculated from normalized attention sources with legacy fallbacks for continuity.
Mentioned in mainstream news or broadsheets — the work has reached beyond academia into public awareness.
Cited in public reference infrastructures (e.g., Wikipedia/Stack Exchange), signalling integration into trusted open knowledge.
Connected to SDG/policy-oriented signals that indicate potential governance and societal relevance.
Present on Amazon, Goodreads, or in patents — indicating consumer interest or industry application.
Foundational dataset reuse reflected in sustained downstream citations.
Dataset appears repeatedly in benchmark-style analytical and methodological usage.
Broad, cross-venue dataset utility across multiple communities and workflows.
Software is repeatedly relied on as part of downstream research methods.
Software supports transparent, repeatable computation and method reuse.
Adoption in open teaching/practice channels and broader community usage.
Every story page uses visual chips and badges in the header to give an immediate read on a work's status and key signals. Hover any badge on a story page for a full explanation of its source and meaning.
Martijn has more than 20 years of experience in scholarly publishing and biomedical research. After completing his PhD in Basel, he built a career at the intersection of publishing, technology, and open science, with a particular focus on books and journals.
In 2014 he was co-founder of Bookmetrix, a Springer–Altmetric initiative that pioneered book- and chapter-level metrics by aggregating multiple signals of reach, attention, and use. Through this work, Martijn gained first-hand experience in the opportunities and limitations of impact reporting — including the importance of transparent data provenance and discipline-sensitive interpretation, especially in the humanities and social sciences.
At openscience.works, he leads partnerships and product direction — building the open infrastructure for a more transparent research ecosystem.