Very interesting example of the complexity of communication. It would be even more useful to see context and concatenations. For example, how many people who extended their thoughts and prayers also spoke up for civil rights?
Look at all nine examples. While any one person might more or less strongly agree with any particular sentiment, none of them are inherently "wrong." You can certainly read a greater inclination for making political points out of a tragedy among some than others.
This reminds me of word clouds (software that digests a body of text and then represents words in size based on their frequency of use. It is fascinating to read a passage and then see a word cloud representation of that passage. But usually it is hard to intuit a particular interpretation solely of frequency of word usage.
Same with this diagram. You think there must be something there, and there probably is, but it is hard to describe exactly what the represented data is telling you.