Dataspace

Dataspace

A sovereign data-sharing platform. Partners find each other’s datasets and train models together, while each dataset stays with its owner.

A sovereign data-sharing platform. Partners find each other’s datasets and train models together, while each dataset stays with its owner.

Request a demo

Cognitive Framework: datasets table with names, IDs, authors, types and creation dates
Cognitive Framework: datasets table with names, IDs, authors, types and creation dates
Cognitive Framework: datasets list with type badges

The most valuable data cannot move

The most valuable data cannot move

Clinical, financial, industrial and public-sector data is locked down by law, contract and trust. It cannot be pooled in one place, so it often stays unused.

Clinical, financial, industrial and public-sector data is locked down by law, contract and trust. It cannot be pooled in one place, so it often stays unused.

Pooling is not allowed

Privacy law and data-residency rules often forbid copying sensitive records to one shared location.

Copies end control

An organization loses control of its data the moment someone else holds a copy.

Datasets stay hidden

Useful datasets sit behind organizational walls. Partners cannot use them, because they do not know they exist.

Partnerships are slow

Each new data partnership becomes a separate legal project that takes months.

Four platform capabilities

Four platform capabilities

Dataspace connects the sites of partner organizations. Each site runs its own catalog, identity, access rules and contract log.

Dataspace connects the sites of partner organizations. Each site runs its own catalog, identity, access rules and contract log.

Shared data catalog

Search datasets across all partner sites from one place. The catalog shows descriptions and terms; the data files stay with their owner.

Owner-defined access

Owners decide which datasets other organizations can find, and define access by role and organization.

Digital contracts

Each use of datasets is covered by a contract that the owning site signs. Contracts expire, and the owner can revoke them.

Federated training

Select datasets from several sites and train one model on them. The data stays at its site.

The promise to partners

The promise to partners

Each partner keeps control of its data, its users and its access rules.

Each partner keeps control of its data, its users and its access rules.

Data stays home

The data files stay in the owner’s storage. Other sites see only the descriptions that the owner chooses to share.

Decentralized identity

Each organization issues credentials to its own users. There is no central credential authority.

Contract log at each site

Each site keeps its own log of the contracts issued for its data and of every change to their status.

Standard metadata

Every site describes its datasets in DCAT-AP 3.0, the European metadata profile for data catalogs.

Request a Dataspace demo

Request a Dataspace demo

See how partners find datasets across sites, receive signed contracts and train one model together.

See how partners find datasets across sites, receive signed contracts and train one model together.

Request a demo