Referring to this example, I want to add that we generally propose to have two separated data ranges/tables for the list of nodes and the list of connections (edges). The shown table mixes two things, the specification of the connections in columns A and B and the specification of node types in columns B and C. This makes it difficult to add a type for department 'Root'. Apart from this, it is fine here, since the graph is a tree and every node appears only once in column B. However, for more general graphs, nodes would appear several times, requiring to duplicate the data of column C for every occurrence.
A more general solution would be to have a sheet for the list of nodes containing the data of columns B and C, but each node only once, and a sheet for the list of connections containing the data from A and B.