Last year a colleague of mine showed me tagcrowd.com, which is a neat way to take a bunch of text and visualize, very roughly, the important concepts in the text. The site performs a frequency analysis of words in the text, and displays it in the intuitive fashion we have all come to know and love when reading blogs or flickr or other tagging sites. However, tag clouds don't have to be about tags, it can work for any mass of text.
Below, I took the titles of all 51 2007 Topcites from SPIRES (as well as abstracts from 37 of them and keywords from DESY for 27 of them) and threw them at the TagCrowd generator to see what it thought. So without further ado, here is an unintelligent, algorithmic, and probably very biased picture of the important concepts in HEP in 2007.
