Mapping CORDIS projects to SDGs

CORDIS – mapping EU-funded research projects to the Sustainable Development Goals

CORDIS provides information on EU-funded research and innovation projects, including project objectives, descriptions, keywords, EuroSciVoc categories, deliverables and other project-related content.

The United Nations Sustainable Development Goals provide a widely used framework for understanding societal, environmental and economic impact. EU Vocabularies provides access to a structured SDG taxonomy, including more specific narrower concepts. The challenge is to use this taxonomy to identify which SDGs are addressed by CORDIS projects and, where possible, to map them to the most precise relevant concepts.

Participants are invited to design a method for linking CORDIS projects to SDG concepts. A project may contribute to more than one SDG. The mapping must therefore support multiple labels and must include an explanation of why each label was assigned.

The solution may use any relevant CORDIS project fields, such as:

• EuroSciVoc categories;
• project keywords;
• objective (coming from the eGrant agreement);
• project description (editorial content);
• project deliverables;
• other available project-related content.

Participants should not only produce a mapping table. They should also analyse and present the results, for example by showing which SDGs are most frequently linked to CORDIS projects, which are less represented, how SDG relevance varies by programme, topic, country, organisation type, project duration or scientific field, or where the mapping is uncertain.

As an optional extension, participants may also compare the SDG mapping with the current European Commission Strategic priorities authority table, where relevant.

Objective

Create a transparent and explainable method to map CORDIS projects to SDG concepts, using the EU Vocabularies SDG taxonomy as the reference framework.

The best submissions will combine sound data analysis, good use of controlled vocabularies, clear reasoning and effective presentation.

Expected output

Participants should submit:

1. A notebook explaining the methodology, data preparation, mapping logic and results.
2. A project-to-SDG mapping table, including:

• CORDIS project identifier;
• project title;
• assigned SDG concept(s);
• narrower SDG concept(s), where applicable;
• confidence score or confidence level;
• short explanation or evidence for each assignment;
• source field(s) used for the assignment.

3. Visualisations showing the main findings.
4. A short discussion of limitations, uncertain cases and possible improvements.
5. A working web-based prototype. 
6. A description document on what you did and why your solution is the best.

Data

Participants are encouraged to use the datasets made available through data.europa.eu, although equivalent information may also be obtained directly from CORDIS. 

Relevant fields may include, but are not limited to:

• project acronym;
• project title;
• objective;
• project description;
• EuroSciVoc category;
• keywords;
• programme;
• topic;
• start and end dates;
• participating organisations;
• countries;
• deliverables and other project outputs, where available.

The reference taxonomy is the Sustainable Development Goals taxonomy in EU Vocabularies:

Sustainable Development Goals

Participants should use the taxonomy structure and prefer the most precise relevant concept. Where a narrower concept is available and applicable, it should be used instead of assigning only the top-level SDG.

Optional reference authority table:

Strategic priority

Task

For each selected CORDIS project, identify one or more relevant SDG concepts.

A good solution should:

• support multiple SDG assignments per project;
• use narrower SDG concepts where possible;
• explain the reasoning behind each assignment;
• distinguish strong, weak and uncertain matches;
• avoid assigning SDGs only on the basis of isolated keywords;
• show how the method deals with ambiguous or cross-cutting projects;
• provide reusable outputs.

Methodological guidance

• Participants are free to choose their own approach for linking CORDIS projects to SDG concepts.
• The proposed method should be transparent, explainable and reproducible. Participants should clearly describe how SDG assignments were generated, how confidence was assessed and how uncertain or ambiguous cases were handled.
• If artificial intelligence, machine learning or other automated methods are used, the solution must explain how the method works and how the results were validated. Black-box mappings without explanation will not be considered sufficient.
• Participants must not process personal data.
• Participants should document any external datasets, tools or models used.

Visualisation

Participants are encouraged to present their findings through appropriate visualisations that help explain the mapping results, highlight patterns in the data and communicate uncertainties where relevant.

Evaluation criteria

Submissions will be evaluated according to the following criteria:

1. Correct use of the SDG taxonomy

The solution uses the EU Vocabularies SDG taxonomy correctly and assigns the most precise relevant concepts where possible.

2. Quality of the mapping

The assigned SDGs are relevant, well justified and not based on superficial keyword matching only.

3. Explainability

The notebook clearly explains why a project was linked to a given SDG. The reasoning is understandable for non-technical users.

4. Reusability

The method and output can be reused or adapted by CORDIS or EDP after the challenge.

5. Data handling

The solution uses CORDIS data appropriately, documents preprocessing steps and handles missing or inconsistent data transparently.

6. Visualisation and communication

The results are presented clearly through tables, charts or other visualisations.

7. Awareness of limitations

The submission identifies uncertain cases, possible biases and limitations of the method.

8. A working web-based prototype

This prototype helps the work of the evaluators.

9. A description document on what you did and why your solution is the best

A short summary of why you think your solution should be awarded.

Submission format

Each submission should include:

• a notebook;
• a mapping table in CSV or Excel format;
• any supporting files needed to reproduce the results;
• a short summary of the approach and main findings;
• a working web-based prototype; 
• a description document explaining why your solution is the best.

Indicative timeline

• Launch of the challenge: 1 August 2026
• Q&A period: until 31 August 2026
• Submission deadline: 30 September 2026, 23:59 CEST (UTC+2)
• Evaluation period: 1–30 October 2026 
• Results announced: 4 November 2026

Awards

The best submissions will receive gold, silver and bronze certificate recognition on the competition platform.

Key terms