In our book the process for Sentiment Analysis is to join a sentiment dataset and do dictionary look ups.

There is a very nice R package called sentimentr that can also be used for sentiment analysis. Check out the website.

In the book different sentiment dictionaries are compared.

What are the sentiment values for each dictionary?

Wordclouds can be used to compare document or compare words sentiment within a document. This can be done for an entire corpus also.

What is wrong with word clouds?

Basically two dimensions are changing when only one variable is changing.