Data Management Advent Calendar
Start with Purpose
Every dataset needs a clear purpose. Purpose clarity leads to better metadata, licensing, and reuse.
Good data management starts with a clear purpose. State why the dataset exists, who will use it, and what questions it should answer. Purpose clarity leads to better metadata, more appropriate licensing, and smoother reuse.
Make EO data truly findable and usable.
Make EO data truly findable and usable.
Apply FAIR from the start: make EO data Findable, Accessible,Interoperable and Reusable. Assign persistent identifiers, use standard metadata, and document access methods so machines and people can both discover and use your data.
A simple DMP prevents tech debt.
GEO data management plan: plan early.
Even a light-weight Data Management Plan helps. Describe what data you will collect or generate, who owns it, how it will be stored,licensed, shared, and preserved. Planning early avoids technical and organisational debt later.
Metadata Is Not Optional
Metadata standards pay off every time.
Metadata is not optional. Use community standards like ISO 19115,INSPIRE or DCAT-AP together with domain vocabularies. Good metadata is the bridge that connects your dataset to users and applications.
Accessibility Beyond ‘Open’
Accessibility needs clarity. Document APIs, access paths, rate limits and examples.
Accessibility is more than saying ‘open’. Document how to obtain the data, which APIs and protocols to use, and any authentication or rate limits. Provide examples so users can quickly test access.
Interoperability for Earth observation
Interoperability thrives on standards: OGC APIs + STAC.
Don’t reinvent, integrate.
Interoperability enables observation integration across platforms and digital twins. Use common formats, coordinate reference systems, and standards such as OGC APIs and STAC. Interoperable data flows are easier to automate and scale.
Context matters.
Share your assumptions, processing steps, calibrations and limitations.
Reusability depends on context. Document provenance, processing chains, algorithms, calibration and known limitations. Clear context allows others to safely reuse your data in new applications.
Respect Indigenous rights and data sovereignty.
Respect Indigenous data rights.
Apply CARE principles; Collective Benefit, Authority, Responsibility, Ethics.
Apply the CARE Principles when data relate to Indigenous lands, peoples or knowledge: Collective Benefit, Authority to Control, Responsibility and Ethics. Co-design data practices with affected communities and respect their rights.
Ethics belong in every EO workflow.
Ethics belong in every EO workflow.
Avoid unintended harm; design with responsibility.
Earth observation data can influence major decisions about land, water and communities. Integrate ethical reflection into the design of data collection, analysis and sharing. Avoid misuse and unintended harm.
TRUST Repositories.
Choose repositories guided by TRUST.
Transparency, Responsibility, User focus, Sustainability, Technology.
Use repositories that follow the TRUST Principles: Transparency, Responsibility, User Focus, Sustainability and Technology. Trustworthy repositories increase confidence, persistence and uptake of EO assets.
Long-Term Preservation
Think beyond projects.
Preserve EO data for decades, not deliverables.
Think of preservation on a 10+ year horizon. Choose durable formats, reliable archival storage, and clear responsibilities. Cloud buckets without a plan are not long-term preservation.
Data sustainability
Make Earth observation services sustainable. Plan governance, funding, maintenance and sunsets.
Data services must be sustainable. Define governance, funding models, maintenance roles and sunset plans. Long-lived Earth observation services need more than project funding.
Licensing Clarity
Licensing should be clear and simple.
Choose CC-BY, CC0 or ODbL – avoid custom terms.
Licensing clarity reduces friction. Use well-known open licenses such as CC-BY, CC0 or ODbL, and state them clearly in metadata and on landing pages. Avoid custom licenses that confuse users.
Show confidence level, not fake precision
Information about uncertainty builds trust.
Share confidence levels, quality flags and error boundaries.
Uncertainty information builds transparency and trust.
Documents and examples matter
Documentation is part of the product.
README files, notebooks, examples — close to the data.
Treat documentation as part of the product, not an afterthought. Provide README files, API docs, notebooks, and usage examples close to the data. Good documentation is a key ingredient of open knowledge.
Version Everything
Version everything — data, code, models.
Help users understand what changed and why.
Everything evolves: datasets, algorithms, models and services. Use versioning for data and code, and describe what changed. Mark deprecated versions clearly so users can migrate safely
Federated Access
Federation beats silos.
Bring computation to the data, not the other way around.
Federated access lets data stay close to its source while remaining usable across systems. Use catalogues, standard APIs and federation patterns instead of building monolithic silos.
Describe all quality dimensions.
Quality is multi-dimensional.
Describe accuracy, completeness, consistency, timeliness and lineage.
Quality is multi-dimensional: accuracy, completeness, consistency, timeliness and lineage all matter. Describe these aspects in your metadata so users can judge fitness for purpose.
User-centric design.
Design with real users early and often.
Data becomes useful only when it meets real needs.
User-centric design means involving users early and often. Talk to cities, agencies, researchers and communities to understand what formats, latencies and interfaces they actually need.
Machine-Actionable Data.
Make data machine-actionable.
Semantic metadata + PIDs = automation-ready EO.
Make data machine-actionable: use structured metadata, standard schemas, persistent identifiers and clear semantics. This enables automation and AI-based discovery and processing.
Persistent Identifiers (PIDs) provide stable and globally unique references for research entities such as people, datasets, and tools. When combined with semantic metadata, they add machine-readable meaning, relationships, and provenance, enabling FAIR data discovery and interoperability, supporting CARE principles through transparent context and attribution, and strengthening TRUST by ensuring traceability, accountability, and responsible reuse across the GEO data ecosystem.
GEO Open Knowledge.
Open Knowledge > Open Data.
Share tools, code, workflows, training and context.
GEO’s Open Knowledge vision goes beyond open data. Share algorithms, software, documentation, training material and context so others can build on your work.
From Data to Knowledge.
Help users bridge the gap from data to insight.
Examples and workflows turn data into decisions.
Help users move from data to knowledge. Provide worked examples, workflows and best practices that show how your data and tools solve real problems.
Stewardship as Community.
Celebrate the stewards behind the data.
Good data is built by communities — honour them.
Good data stewardship is a shared practice. Acknowledge contributors, credit communities, and share lessons learned. Celebrate open, responsible data management as part of Earth intelligence for all.