Testing hierarchical clustering

Back in 2013 we developed a simple tool for helping Antonio Gutierrez RubĂ­ to explain the benefits of Open Data and Transparency to Spanish politicians.
The tool was called MAPO, and you can see it in action here. Several months later, I was testing some hierarchical cluster techniques for this set of text files and I came up with this (as far as I remember, this is the best clustering output that I never had). Clusters in the image simply transmit the set of words which are really linked both in the text and in the historical contextual situation.