Federation

Federation is a core feature of the data network we are building. Using it is a good way to enrich your own data and it also gives a new way for others to discover and reuse them.

Federation simple schema

Federation allows to redistribute, in full or in part, an open dataset from your domain (OpenDataSoft instance) or from another domain. It is a way to collect external sources of data on its own domain.

Source

To federate a dataset, simply go to the Catalog > Dataset page and hit New dataset. Then, hit the little arrow on the right side of the Add a source button.

Federation source

There is then a link Add a dataset from the OpenDataSoft network, it leads to a console from where it is possible to configure a new federated source.

Federation source config

Click on Select a source dataset.

Federation sources

Different sources are available :

  • The same domain
  • All open datasets from every OpenDataSoft domains
  • Domain’s subdomains (optionnal)

From there the next step is to either select a whole dataset or to select a part of it. The filters are the same as in the dataset explore console, there are pre-defined filters, a query box and, if available, the geofilter.

Federation filter source

Once the data is selected the dataset appears in the backoffice like a normal one, except that there is no Processing tab. Indeed, the data is not duplicated and it is not possible to transform it. Only the metadata and visualization configuration can be overriden.

Federation federated source

Overriding metadata and visualizations

Both metadata and visualization can be overriden. Simply hit Override and fill the value you want.

Federation override meta

It is possible to come back to the original value by hitting Return to original value.

Federation override meta

Limitations

The federation does not duplicate the data. Only the metadata and visualizations are duplicated, hence can be overriden. There are two reasons for that:

  • It is then possible to redistribute data without impacting the size of the data defined in the contract,
  • When the data changes, the federated dataset is always up-to-date.
Federation schema

Important

Currently, when there is a modification in the original dataset related to the dataset schema, metadata or visualizations configuration, the federated dataset is not automatically updated. You’ll then have to unpublish / republish your federated dataset to reflect the changes.