The Open Data Directive is Directive (EU) 2019/1024 on open data and the re-use of public sector information, the EU law that requires public sector bodies to make the documents and data they already hold available for re-use for commercial and non-commercial purposes. Within EU data regulation it is the oldest pillar, a recast of the PSI Directive of 2003, and it sits beside the Data Governance Act and the Data Act in the wider framework of tech regulation.
The Directive does not create a right to obtain information. It governs what happens after data is accessible: the formats, the charges, the licenses and the conditions under which companies, researchers, journalists and citizens may reuse it. Member States had to transpose it into national law by 17 July 2021.
How the Open Data Directive works
The Open Data Directive works by turning re-use into an obligation for information that national access rules already allow to be disclosed. When a document is publicly accessible, the holder must permit its re-use, process requests within fixed deadlines and avoid discriminatory or exclusive conditions.
The text follows the principle of open by design and by default. Data should be published in open, machine-readable formats together with its metadata, ideally through public portals, so that re-users do not depend on scanned documents or proprietary file types. For data that changes frequently, such as traffic sensors or air quality readings, the Directive requires dynamic data to be provided immediately after collection through application programming interfaces (APIs), with a limited delay allowed only where immediate release would exceed the technical or financial capacity of the body.
Individual requests must be answered within 20 working days, extendable by another 20 working days for extensive or complex requests. A refusal must state its reasons and the available means of redress.
Who and what the Directive covers
The scope of the Open Data Directive is broader than its predecessor because the recast extended it to public undertakings and research data. The table below summarizes the main categories of holders and how strongly each is bound.
| Holder | What is covered | Strength of obligation |
|---|---|---|
| Public sector bodies | Documents held within their public task | Full re-use obligation |
| Public undertakings | Utilities and transport data from services of general interest | Re-use rules apply only to data they choose to make available |
| Research performing organizations | Publicly funded research data already made public in repositories | Must be reusable; open access policies set by Member States |
| Libraries, museums and archives | Cultural heritage holdings | Covered, with special rules on charging and exclusivity |
| Schools and other cultural bodies | Educational and performing arts material | Excluded from scope |
Several categories are excluded entirely: documents protected by third-party intellectual property rights, data excluded from access for reasons of national security, statistical or commercial confidentiality, and documents whose re-use would conflict with personal data protection. These exclusions explain why some publicly visible records remain off limits for re-use.
Charging, licensing and exclusive arrangements
The default rule is that re-use is free of charge, or at most limited to the marginal costs of reproduction, provision, dissemination and anonymization. Above-marginal pricing is allowed only for defined exceptions, and even then charges must not exceed costs plus a reasonable return on investment.
| Holder or data | Maximum charge | Condition |
|---|---|---|
| Most public sector bodies | Marginal cost or free | Standard rule |
| Bodies required to generate revenue | Cost recovery plus reasonable return | Revenue duty must be set in law |
| Libraries, museums and archives | Cost recovery plus reasonable return | Includes preservation and rights clearance |
| Public undertakings | Cost recovery plus reasonable return | Applies to voluntarily released data |
| High-value datasets | Free of charge | Limited exceptions for some bodies |
| Research data | Free of charge | Covered by the free default |
Licenses must be objective, proportionate and non-discriminatory, and Member States are encouraged to use standard open licenses so that re-users do not negotiate terms dataset by dataset. Exclusive arrangements that grant one company privileged access are prohibited in principle. An exception for a service of general interest must be reviewed at least every three years, and exclusivity for the digitization of cultural resources is limited to ten years as a general rule.
High-value datasets
High-value datasets are the core innovation of the recast: categories of public data whose re-use is expected to bring the greatest economic and social benefit, and which must be published free of charge, in machine-readable form, via APIs and as bulk downloads. The Directive names six thematic categories, and Implementing Regulation (EU) 2023/138 lists the specific datasets, applicable from 9 June 2024.
| Category | Typical datasets | Typical re-use |
|---|---|---|
| Geospatial | Administrative units, addresses, cadastral parcels | Mapping, logistics, property services |
| Earth observation and environment | Emissions, water, nature protection | Climate analytics, compliance tools |
| Meteorological | Observations, forecasts, climate data | Weather apps, agriculture, energy trading |
| Statistics | Population, labor, economic indicators | Research, market analysis |
| Companies and company ownership | Basic company register data | Due diligence, anti-fraud checks |
| Mobility | Transport networks, inland waterways | Route planning, mobility services |
The high-value dataset rules have a separate page of their own because they carry detailed technical requirements on formats, metadata and API documentation. For the Directive, the key point is that this list is dynamic: the Commission can add new datasets through further implementing acts.
How it compares with related EU data laws
The Open Data Directive is often confused with newer EU data laws, but each addresses a different layer of data sharing. The comparison below shows where the boundaries lie.
| Instrument | Main subject | Relationship to the Directive |
|---|---|---|
| Open Data Directive | Re-use of openly accessible public data | Baseline for public data re-use |
| Data Governance Act | Re-use of protected public data, data intermediaries, data altruism | Covers data the Directive excludes, such as confidential data |
| Data Act | Access to connected product data and business-to-government requests | Mainly private sector data |
| GDPR | Processing of personal data | Re-use of personal data must still comply |
| INSPIRE Directive | Spatial data infrastructure | Supplies much of the geospatial data |
A practical rule of thumb: if public data is openly accessible, the Open Data Directive applies; if it is protected but could be shared under safeguards, the Data Governance Act sets the conditions; if it is held by private companies, the Data Act is the more likely source of obligations.
Limitations and common misconceptions
- It is not a freedom of information law. Whether a document can be accessed at all is decided by national access rules.
- Personal data is not freed by it. Anonymization or a lawful basis under the GDPR remains necessary, and the Directive explicitly yields to data protection law.
- Public undertakings are not forced to publish. Their obligations attach only to data they decide to release.
- Quality varies by country. Transposition was uneven, and portal coverage, metadata quality and API availability still differ widely between Member States.
- Free does not mean without conditions. Licenses may still require attribution or indicate changes, as long as conditions remain proportionate.
The idea in brief
Governments collect huge amounts of information while doing their jobs. Think of maps, weather records, company registers, bus timetables and statistics. The Open Data Directive says that once this information is public, anyone may reuse it. Usually this is free of charge, in files that computers can read easily.
Picture a public library that lets you copy its books and build something new from them. That is the spirit of the rule. The main takeaway is simple: public data paid for by taxpayers should flow back to society. It becomes raw material for new services, research and better decisions.
A practical scenario
A small startup in a mid-sized EU country wants to build an app that warns delivery drivers about road closures and bad weather. It needs the national road network, real-time traffic sensor readings and weather forecasts, all held by public bodies.
Under the Open Data Directive, the transport authority publishes its road network as a high-value dataset through a documented API at no cost, and the sensor feed is provided as dynamic data within minutes of collection. The meteorological office offers forecasts under a standard open license that only requires attribution. When the startup asks the city for additional parking data, the request is answered within 20 working days.
The result is a working product built in weeks rather than months, with no licensing fees and no exclusive deal that would have favored a larger competitor.