Add ability for users to view token count and add custom stop words before model building

[taalbrecht]

Desired features:

1. Display tokens by frequency (per document or overall count ideally)
2. Display discovered multi-word phrases and allow users to remove or add to the list
3. Allow users to add or remove tokens from the stop word list