Datenlotse

Datenlotse takes public bodies from a map of their data flows into the operational work of publishing open data. It answers the follow-on question: what now, in a way that is legally sound, standards-compliant and in the right order? Like all these tools it runs in the browser, with no account and no server.

Once it is clear which data the organisation holds, the harder work begins. Which holdings can be published? Who is responsible? Which licence fits? What has to be checked first?

Datenlotse arranges these questions along five phases: taking stock, governance and roles, a data inventory following DCAT-AP.de, risk assessment, and pseudonymisation. Each building block shows where you stand and what the sensible next step is.

You start with a maturity check across seven dimensions, grounded in established models such as ODRA, the EU Open Data Maturity Assessment and DCAT-AP.de. The sources are shown in the interface. The result is a score with a prioritised next step.

You then import a map of your data flows, from DatenGraf or as your own CSV in the same schema. Dataset candidates are derived from it for you to complete. A quality check flags missing mandatory fields, licences outside the Open Definition and invalid entries. A licence guide leads through two questions to one of the four most reusable licences.

A traffic-light clearing view rates each dataset for risk. The logic is a published decision tree, not a model: green is uncontested, red is blocked, and your review capacity concentrates on amber. For free text there is a redaction function that recognises eleven entity types from the German administrative context and supplies a mapping table.

What comes out: a DCAT-AP.de catalogue as JSON-LD, inventory and clearing as CSV and PDF, a RACI matrix, and template documents such as a release form, a processing-record extract and a draft policy.

Datenlotse publishes nothing. It prepares publication. Automated updating and the technical connection to a portal are not included; they depend too heavily on the specific system landscape to be covered generically.

The redaction finds only what its patterns describe. Manual review remains obligatory. And none of the assessments is legal advice: responsibility stays with the organisation.

Without a map of the data flows there is no basis. The tool starts where such a map exists.

From the same workshop practice as DatenGraf. Once the map is drawn, every second organisation asks the same thing: and now? That called for something ready to use straight away, costing nothing, raising no data protection questions and usable without an introduction.

It is open because the work carries on after the session, usually with the organisation's own fields and vocabulary. The source code is public on GitHub, under an MIT licence.