Typical usage of the output data is to load into Excel.
Excel is fine for this, but overkill when all that is needed is sorting, filtering, and searching.
The overkill isn't just in terms of price, but also UI complexity, executable size, etc...

A custom UI could be better.

Also mine strings from classes for analysis like the unix strings command mines them from binaries.

Allow for the option of starting with a directory of jars rather than requiring jdeps to be run separately,
Perhaps even all for starting with an EAR or WAR.

When a large number of uses or names are specified for analysis, the resulting table can be hard to use due to the column count.
Some form of column analysis is needed to help reduce the columns when possible.

What is the correlation between any two columns?
Are there columns that are functionally aliases?