The Barcamp Open Science (commonly #oscibar) is an annual “unconference” open to all who are interested in Open Science—researchers, practitioners, novices, and experts alike. Its defining feature is its participant-driven format: attendees propose, vote on, and lead sessions on any aspect of Open Science, facilitating lively and inclusive discussions. The event is organised by a team of volunteers with support from Leipniz Open Science and welcomed 45 participants at Wikimedia Deutschland in Berlin on June 18 2025. Since its inception in 2015, the Barcamp has fostered community-led exchange around Open Science. Over the years, it has addressed a growing range of topics—from Open Access, Open Data, and persistent identifiers to multilingualism, equity in research, reproducibility, non-STEM practices for openness, and resilience. KOMET team member Daniel Nüst co-organised this year’s event and also contributed a session on geospatial metadata in scholarly communication. Here is his report. A link to the reports from other sessions will be added here soon.
Geospatial metadata—information about the where and when of research—has the power to connect knowledge across disciplines in exciting new ways. Imagine being able to ask:
- Which places are studied in a particular research paper?
- Where were the participants in a social science study interviewed?
- When and where did an archaeological excavation take place?
- Which parts of the Global South haven’t been studied for ecosystem services?
These kinds of questions sound simple, but the current infrastructure of academic publishing often doesn’t support them well. Research articles, data sets, and digital collections are rarely linked to geographic information in a way that is easy to use or search. Journal websites, data repositories, and metadata standards typically don’t “put research on the map.” Even when location is important to the study, it often isn’t included in a structured, searchable way, but hidden in the full text, supplementary materials, or metadata for data.
In this session, we explored this topic from many perspectives. The session was initiated by Daniel Nüst, a research software engineer working on projects (OPTIMETA, KOMET) that aim to fix this gap. He introduced some of the first tools that bring geospatial metadata into academic publishing. One of these is geoMetadata, a plugin for the widely-used journal platform Open Journal Systems (OJS), which helps journals capture and display geographic metadata for research articles. Another is OPTIMAP, a discovery website that visualizes research outputs on a map. Besides geospatial metadata, the mentioned projects also facilitate using citations and persistent identifiers (PIDs), such as ROR or IGSN, to enhance the scholarly metadata commons.
Since several participants had technical backgrounds, the conversation began with practical details. Daniel explained that OPTIMAP’s technical side is quite straightforward: rather than needing precise survey-grade coordinates, broad latitude and longitude information is enough for the use case of research discovery. The data is stored in PostgreSQL using PostGIS, a standard spatial database extension. However, things get tricky when connecting the OPTIMAP dataset to services like Wikidata, which can only represent point locations and often needs multiple entries for something as basic as a country’s extent.
The vision for OPTIMAP goes beyond just individual articles: it’s about building a map filled with research from many different sources. So how do we get those contributions onto the map? Participants had creative ideas:
- Use large language models (LLMs) to extract location references from text
- Pull coordinates from linked data (e.g. DataCite records with geospatial metadata)
- Use citations of physical samples with known locations (like IGSNs)
- Extract place names (like "Berlin") used as tags or keywords (e.g., using gazetteers)
Daniel highlighted that some repositories already do this well—such as PANGAEA and GFZ Data Services—and that these domain-specific repositories could serve as models for generic research data repositories. The group also explored entirely new use cases: for example, enabling museum visitors to explore artifacts by places of origin, or supporting citizen science projects based on users' locations. Developers in the room proposed improving standards to better support spatial metadata and encourage wider adoption.
One challenge Daniel faces is finding journal partners to adopt the OJS plugin. Many independent journals have limited time and technical resources. When asked if the plugin could be made useful for larger publishers, Daniel explained that while some parts could be reused, each publishing system is different. His team is also working on a version for Janeway. Ultimately, the technical setup is manageable—it’s the user experience and communication that matter most.
Even basic terms like “overlap,” “within,” or “touch” (used to describe spatial relationships) can confuse users if not carefully explained. This highlights a key insight from the session: communication is crucial. The term geospatial metadata isn’t widely understood outside of certain technical or Earth sciences communities. But the benefits could be greatest for disciplines that don’t typically think in terms of maps or coordinates—because that’s where new and surprising research connections can emerge.
Daniel left the session with new ideas for applying geospatial metadata in fields like:
- Metascience (studying science itself)
- Diversity and inclusion studies
- Research equity, to explore geographical biases in what gets studied
- And importantly, while political borders and names can be contested or confusing across languages and jurisdictions, simple coordinates are not.
The discussion ended by returning to the question of data collection. The group explored how to assist researchers in creating spatial metadata—either through smart suggestions (powered by machine learning), or simple tools. For example, a system could help the users with:
- Retrieve and suggest the shape of a region from OpenStreetMap using the Overpass API
- Suggest location names from free-text descriptions (NLP)
- Let users draw a rough shape on a map with a “thick marker” to define a location, track/line (e.g., river), or area
The key is keeping it simple and user-friendly. The goal isn’t to create perfectly precise maps, but to help people put their research on a map—literally. Daniel extends his thanks to all participants for their thoughtful questions and ideas. The session reinforced that while the core concept is broadly supported, it needs clear value propositions and real-world examples to reach communities beyond the traditional geospatial domain. User perspectives must remain central as the work continues. Get in touch with Daniel if you are interested in the topic!