If you have data that you want to use with the CMS (and also manage), it is possible to initially work directly with JackRabbit and once the data has been imported and tested, import into the CMS repository. You can work with regular nt:unstructured nodes, and import them into Magnolia and build standard dialogs to manage the properties of the nodes. You can export the root node you wish to import into magnolia and then include the file using standard naming convention under the config/resource/mgnl-bootstrap-samples/
One point to keep in mind is that Magnolia uses a child MetaData node to track changes and activation status of nodes. If you plan on managing the imported nodes using Magnolia and wish to see the status of your changes and activation, you will need to add a MetaData child node to each node you wish to manage.
Note that Magnolia provides a data module for dealing with data that you wish to import from external sources. We chose not to use this, since we wished to keep our code as generic as possible. The entire books repository we created with all our data is pure JCR with no Magnolia dependencies. This makes it possible for us to use the same data and associated model and repository classes with any other JCR compatible CMS.