What happens when an Open Science infrastructure and a research infrastructure build a monitoring system together, not simply to report on research outputs, but to understand how those outputs connect across publications, data, and software?
The collaboration between SciLifeLab and OpenAIRE shows how co-creation can turn fragmented information into a more connected, usable view of research, helping institutions better understand, support, and advance Open Science in practice.
Why we built this together
SciLifeLab needed a way to better understand its research landscape across publications, datasets, software, and related outputs, and to connect that view to its Open Science policies, data-sharing ambitions, and community guidance.
Before this work, these signals were distributed across multiple systems, with inconsistent links between research objects and reliance on manual updates. OpenAIRE brought an existing large-scale scholarly knowledge graph, services for searching, exploring, and monitoring research outputs, and enrichment pipelines that could be adapted to SciLifeLab’s context.
This need also sits within a broader national context, where Sweden has set a goal that publicly funded research should be fully open and accessible by 2026. Meeting that ambition requires not only enabling Open Science practices, but also being able to understand and follow how those practices take shape across institutions.
Together, the goal was not only to count outputs, but to make Open Science at SciLifeLab more measurable, actionable, and most of all, visible.
Co-creation in Practice: How decisions were made
This work was not a simple handover from provider to client. SciLifeLab brought the institutional questions, curated publication knowledge, infrastructure-unit context, repository knowledge, policy priorities, and critical validation feedback. OpenAIRE brought the OpenAIRE Graph, CONNECT and MONITOR services, metadata aggregation, affiliation logic, dashboarding, and full-text mining pipelines for datasets, software, funders, and Data Availability Statements (DAS).
The work evolved through repeated cycles: understanding SciLifeLab’s needs, deploying a first version, validating together, identifying what information was missing or misleading, improving the data and rules, and then releasing and refining again.
This iterative loop is visible throughout the roadmap, meeting minutes, and coverage analysis, and made it clear that meaningful deployment relied on sustained engagement from SciLifeLab: validating outputs, interpreting results, and shaping the service so that it reflected the realities of its institutional environment.
The Journey: Milestones that show collaboration
The collaboration began through two linked tender tracks: one focused on scientific impact monitoring and one on data and software mining. OpenAIRE’s implementation plan combined them into a single progressive roadmap: deploy a SciLifeLab-branded instance, integrate mining outputs, release enriched indicators, refine dashboards based on feedback, and provide training and ongoing support.
From the outset, the process called for more than the delivery of a ready-made service. It involved a sequence of steps on SciLifeLab’s side as well: validating coverage, reviewing results, configuring the CONNECT environment, and helping define how the service should appear and function within its institutional setting.
This work included coverage checks against SciLifeLab’s own curated sources, validation of organisational affiliations, and configuration of the CONNECT instance through visual identity choices, introductory texts, images, FAQs, and other contextual elements that helped make the service intelligible and useful for its intended community. What emerged was therefore a service shaped through stages of development, configuration, and refinement, rather than something presented as finished from the outset.
The platform evolved through successive iterations, in which new data sources, indicators, and mining capabilities were tested and refined together with the SciLifeLab team. The first version was released in September 2025, followed by several months of iterative work on coverage, affiliations, dashboard behaviour, funders, and mining outputs. The SciLifeLab Open Science Monitor service was officially launched in February 2026, with further enrichments continuing as part of the collaboration rather than being treated as post-project extras.
What We Co-created: The platform in plain language
The result of all this work is a SciLifeLab discovery and monitoring environment built on OpenAIRE CONNECT and MONITOR. It allows users to,
- Discover SciLifeLab-related research outputs across publications, datasets, software, and other products
- Explore links between outputs
- Monitor Open Science indicators
- Use those insights to inform guidance, curation, and future improvements.
What makes it unique is not just the interface, but the fact that it is tailored to SciLifeLab’s structure and needs, transparent about what is and is not yet covered, and designed to improve as new evidence and feedback arrive. That degree of customisation grew out of an extended process of exchange, through which priorities were clarified, representation was adjusted, and the service became progressively aligned with the needs of the SciLifeLab community.
The Hardest Part: Data coverage and why co-creation mattered
One of the most revealing moments in the collaboration came when the first version of the platform was compared with SciLifeLab’s own curated publication lists. That analysis showed both strengths and gaps. Many outputs were already visible in the OpenAIRE Graph, but the overlap was only partial. Some publications were too recent to be present yet; many outputs were in OpenAIRE but not yet recognised as SciLifeLab; some relevant research products surfaced through the OpenAIRE Graph had not yet been incorporated into SciLifeLab’s curated publication records; and some records appeared due to imperfect affiliation matching or historical anomalies. Additionally, many outputs were not mined by the OpenAIRE Graph due to incorrect affiliations provided by the authors.
Rather than undermine confidence in the platform, this comparison made the collaboration stronger.
It functioned as a reconciliation exercise between curated institutional knowledge and large-scale aggregated metadata, helping both sides better understand where records aligned, where they diverged, and where additional curation or refinement was needed. It created a shared baseline, identified where affiliation logic and mining needed refinement, and made it possible to define concrete improvement actions around organisational identity, missing links, false positives, pre-2013 filtering, and curated DOI integration. Moreover, it drew attention within SciLifeLab to the correct affiliation of research products and opened a country-wide project for establishing consistent affiliation standards and guidelines for all institutions connected to SciLifeLab.
Mining: Turning full text into better links
A major added value of the collaboration lies in full-text mining. OpenAIRE already had mining workflows for linking publications to projects, funders, datasets, and software, but SciLifeLab’s needs helped push these capabilities into a more visible monitoring context. In particular, the collaboration accelerated work on DAS detection and extraction, as well as improved identification of dataset and software references within publication full text.
Early evaluations of the updated extraction module on a large external dataset showed high precision, and the SciLifeLab corpus is now being processed at scale, with over twenty thousand candidate statements already identified and a post-processing phase underway to remove false positives caused by PDF-to-text conversion artefacts. The goal is to transform signals buried in the full text of publications into structured links that improve the visibility of datasets, software, and research outputs connected to SciLifeLab.
What We Learned: Transparency, interpretation, and responsible use
One lesson from this work is that an Open Science monitor is most useful when it is honest about uncertainty. A platform like this can already support discovery, indicator tracking, and better visibility of connections between outputs, but it should not be treated as a perfect or final representation of reality. Some signals depend on metadata quality, some on affiliation practices, some on full-text availability, and some on choices that only the community itself can clarify. That is why responsible interpretation matters: the monitor should help SciLifeLab learn where practices and data flows are working well, where they need support, and where further curation or mining improvements are still needed.
What’s Next: Co-creation continues
The launch is not the end of the journey. The next steps already identified in the collaboration include continued refinement of SciLifeLab-specific affiliation logic, stronger coverage of funders through SWECRIS and Crossref-linked project data, post-processing and integration of DAS results, ongoing dashboard and indicator refinement, and further improvement of links to datasets, software, and infrastructure units. In that sense, the platform is not just a reporting tool. It is also a shared workspace for strengthening the visibility and interpretation of Open Science practices over time. What the collaboration has shown is that a service of this kind gains robustness and relevance through continued exchange, especially where validation, contextualisation, and local configuration remain essential to its development.
This wider exchange is also supported through OpenAIRE CONNECT Community Calls, which bring together Community Gateway Managers, as well as those interested in taking on such a role, to learn about new developments in CONNECT, share feedback across communities, and discuss features they would like to see evolve. In this sense, co-creation does not stop at the institutional level; it is also sustained through a broader community of practice around gateway development and use. Further information on upcoming and previous community calls is available on the OpenAIRE CONNECT Community Calls page.
