The Internet of (frankly a bit weird) things....
The good thing about a data-driven automaton is that it creates data by-products. This tag cloud is built from the auto-saved word log created by the Psychic Hive Mind Fortune Teller. It is formed from around 4,000-5,000 distinct terms that it was reading from the Twitter conversation happening live during MuseumNext 2013 Amsterdam. The log here was built up by real-time monitoring over the first day of the conference (Monday 13 May 2013)
This log is what it used to construct fortune-readings. Actually, this is the cumulative dataset, that built up over the whole day. On the day, at any one time during the conference, the Fortune Teller has the words from the last 100 tweets in its brain.
This is probably about 600-700 words, not including any hashtags, usernames or urls. These were stripped out. Some stop words have also been removed too, like 'the' and so on.
Specimen tag cloud for MuseumNext13 May 2013
The Fortune Teller is a physical object powered by a Twitter application running as a Processing sketch. It was monitoring the content of tweets from the collective discussion using the Twitter search: twitter.com/search?q=museumnext
Credit: This tag cloud was generated using the rather nice (and still slightly beta) TagCrowd at http://tagcrowd.com/