Some of the graphs were based on keywords selected by the editors, such as for example this graph:
Looking at it, you may have wondered: “but what about this other word?”
Interactive version of the circle plot
Wonder no more! Now, it’s easy to find out through an interactive version of the same graph. You can insert comma-separated words to create your own circle plot based on #SOTEU speeches. As for the original graph, common stopwords (e.g. “and”, “or”, etc.) and self-referential terms (“european”, “commission”, etc.) have been excluded. Terms should all be lower-case, and to facilitate meaningful comparisons this interactive defaults to *term*, so the results of “example” would correspond to the the sum of occurrences of “example” and “examples”. 1
The final version of the graphs published included a number of wordclouds, but not a type of wordcloud that I thought was particularly interesting. A wordcloud based on all the words found in the same sentence as a given keyword. For example, what words are found in the same sentence as “crisis”?
But hey, what if you’re interested in some other keywords? There you go. You may want to try russia, balkan, or whatever you are interested in.
This is mostly what you want, but beware, in particular with short words: “eco” would present the sum of all occurrences of “economic”, “ecologic”, etc.↩