Open Knowledge Maps
Open Knowledge Maps is an AI-based tool designed to enhance the visibility and accessibility of scientific knowledge for both science and society. It is availab...
Last verified:
What is Open Knowledge Maps?
Open Knowledge Maps is the world's largest AI-based search engine for scientific knowledge that creates visual knowledge maps to provide instant overviews of research topics. The tool uses artificial intelligence with natural language processing to aggregate and display publications according to topic similarity, presenting users with a topical overview based on the most relevant resources matching their query. It dramatically increases the visibility of research findings for science and society alike.
Key features include visual knowledge maps showing main topic areas as bubbles with related papers and concepts attached to each area, the ability to choose between BASE (Bielefeld Academic Search Engine) or PubMed as data sources, AI-generated labels for research areas using relevant concepts, filters to refine results by year, language, or content type, and the ability to save, export, or share knowledge maps. The platform highlights open access resources, allowing users to read most content directly from within the interface with fulltext only a click away if not immediately accessible.
Open Knowledge Maps is designed for researchers, students, librarians, academics, and anyone interested in exploring scientific literature efficiently. It is particularly valuable for those new to a research field who need to learn the
Open Knowledge Maps pricing
Pricing model: Free
Open Knowledge Maps is completely free for individual users with no cost to use the core visual search engine. The platform operates on a sustainable, community-supported model as a charitable non-profit organization. Organizations can become supporting members through membership tiers: Basic Membership at EUR 480/year, Sustaining Membership at EUR 780/year, and Visionary Membership at EUR 1280/year. Supporting members gain access to exclusive features including Custom Services for embedding OKMaps components in their own discovery systems and VisConnect for tailored technology integration. All services are free and openly licensed with no licence fees and no lock-in effects for individual users.
Open Knowledge Maps pros
- Completely free for individual users
- Visual knowledge maps provide instant topic overviews
- AI clusters similar resources together for easier navigation
- Identifies relevant concepts and labels research areas automatically
- Highlights open access resources prominently
- Most content readable directly within the interface
- Access to millions of research papers from PubMed and BASE
- Works for research topics in any discipline
- No advertising or commercial bias as non-profit
- Open source visualization framework available on Github
- Filters available by year, language, and content type
- Can save, export, and share knowledge maps
- Citation information provided for each map
- Trusted data providers ensure quality academic content
- Custom services available for institutional integration
- Training materials and workshop kits available for communities
- Multilingual support with area titles in different languages
Open Knowledge Maps cons
- Only uses top 100 resources to create each map
- Up to 30 seconds loading time to create a knowledge map
- Important resources may be missing due to 100-resource limit
- Not optimized for natural language prompts, keywords work better
- Visualization quality depends on metadata quality
- Some area titles may be multilingual due to poor language metadata
- No own index built yet, relying on external APIs
- Limited citation metrics available in BASE integration
- Significant development effort needed for roadmap items seeking funding
Frequently asked questions about Open Knowledge Maps
How do you define 'most relevant' when talking about most relevant resources?
Open Knowledge Maps uses the relevance ranking provided by either the PubMed API or the BASE API, depending on your choice. Both mainly use text similarity between your query and the article metadata to determine relevance. PubMed has a detailed description of their relevance ranking, while BASE uses Lucene (via Solr) for their ranking system.
Why are you only using the top 100 resources to create the map?
The tool keeps resources at a manageable amount since 100 resources are already 10 times more content than presented on a standard search results page. There are cognitive considerations to minimize cognitive load, plus technical limitations as most data providers impose limits on retrievable items or slow down larger queries. Building their own index to provide more content is on their roadmap but requires funding.
Why does it take up to 30 seconds to create a knowledge map?
Each query is processed live: first sent to the selected data provider (PubMed or BASE) which returns the most relevant results in around 15 seconds, then the AI pipeline analyzes this data to create the knowledge map in another 15 seconds. Reducing loading times would require building their own index, which is on their roadmap but needs funding due to significant development and maintenance effort.
How do you ensure the quality of research included in a knowledge map?
Open Knowledge Maps uses only trusted data providers such as BASE, PubMed, and OpenAIRE. These providers carefully review data sources to ensure they include only academic content. However, even with careful review, mistakes can occur, and users can contact [email protected] if they find content deemed unscientific.
Are the maps generated based on full text analysis or metadata analysis?
Resource grouping is based on article metadata, specifically titles, abstracts, authors, journals, and subject keywords to create a word co-occurrence matrix between articles. Clustering and ordination algorithms are performed on this matrix. Area labels are generated from subject keywords of articles in each area, approximated from abstract and title when subject keywords are missing.
How are the area titles created?
Area titles are created from subject keywords of resources assigned to the same area. The system collects single keywords and phrases up to three words long, scores them according to frequency in one area compared to others using a TF-IDF algorithm for extractive summarization, and selects the top three keywords or phrases for each area. When subject keywords are missing, they are inferred from titles and abstracts.
What does the placement of areas (bubbles) and resources mean?
Closeness of areas implies subject similarity - the closer two areas, the closer they are subject-wise. Overlap implies strong subject similarity but doesn't mean areas share common resources since resources are assigned to single areas only. Centrality implies subject similarity with the rest of the map, not importance. Resource placement within an area has no specific meaning as they are moved during initial arrangement to avoid overlap.
What determines the size of areas and resources?
In PubMed integration, paper size is determined by citation count - more citations means larger size. Area size is the sum of citations from resources in that area, with metrics from Crossref. In BASE integration, area size is determined by number of resources it contains - more resources means larger area. All BASE resources are the same size since no additional metrics are available.
How should I cite Open Knowledge Maps?
To cite an individual map, use the citation provided by clicking the cite button on the left side of each map. To cite the open source software Head Start, see the read-me on Github which includes relevant research resources. To reference the website and search, use: Open Knowledge Maps (2019). Open Knowledge Maps: A Visual Interface to the World's Scientific Knowledge. https://openknowledgemaps.org
How is Open Knowledge Maps funded?
Open Knowledge Maps is a charitable non-profit organization run by dedicated team members and volunteers, funded through collective effort. Organizations become supporting members to co-create the platform. They also seek third-party funding for roadmap items. Individuals can help sustain the platform by making donations. Contact [email protected] for supporting membership or [email protected] for funding specific efforts.