ChoralTech: Use Word Clouds to Interact with Text

Cross-posted at
Whether implicitly or explicitly, the pieces that we select for our concerts tell a story and convey meaning. As conductors, choosing areas of text to emphasize or using specific diction can be part of our toolkit to communicate our interpretation of the piece. That interpretation has to be relatively unified, though– while we may encourage our musicians to each bring something personal to their performance of a work, the ensemble still needs to perform the conductor’s intention: “Everyone emphasize your choice for the most important word in this line” is a fun rehearsal strategy but probably shouldn’t happen in the concert.
You can recruit and honor your musician’s individual interpretations using visuals, however. A word cloud is a graphic which is sourced from a piece of text. The program that you use to generate the word cloud will display the most commonly-occuring words largest, allowing the viewer to quickly see what words seem to be the most important out of the selection. Simply inputting the text of the work may or may not result in anything interesting (Whitacre’s “hope, faith, life, love” would give you four words of the same size, for example). The fun starts when you open up the word cloud to crowd submission. By asking your singers to generate the text for the word cloud based on their interpretations of the work, you can create a graphic which conveys the group’s individual reactions and shows the most common submissions.
Some examples of questions that might get this started:
  • In your opinion, which is the most (/are the 3 most) important word(s) in this piece?
  • When you listen to this piece, what emotions are the strongest for you?
  • What do you imagine or visualize while performing this piece?
Anything which generates a list of ideas will work. Remember that the software is looking for word matches, so steer towards descriptive words instead of long sentences or phrases. Once you have your list from your singers, you can upload a text or Word file to a web site such as Wordle or Tagxedo, or type the text directly into the webpage to get your graphic. Either site will give you an image file which you can then use on a web page, print in your programs, or project on the wall in the concert to create a visual aspect to your performance.
While you can take time in a rehearsal to generate the words, create the word cloud at home and bring it back to the next rehearsal, you can also generate them live with a little Google Docs-fu. For an interactive experience for your audience, imagine how you could have the audience generate one live as the musicians sing (or in between pieces/ensembles, for example). If the following instructions scare you, find a techie friend or singer to help you put it together: it’s actually not that complicated. If you construct a Google Doc which is open to the public, and feed the URL of that document to either Wordle or Tagxedo, the word cloud will generate off of whatever words have been added to the Google Doc. It’s not live, so you actually have to have someone refresh the word cloud image to pull in words as they’re being added. You could also set up an auto-refresh of your browser to do it for you. Then, it’s just a matter of connecting your audience to the Google Doc: projecting a QR code on the wall or providing a TinyURL would work for those with smartphones. For true power users, use IFTTT to create a recipe which would add texts or Tweets sent by the audience to the Google Doc.
Poll Everywhere has a free version of polling which will generate a word cloud, but the free version is rather limited in the number of people who can respond, so it’s unlikely to be useful for a large group of performance situation.
However you generate the text, though, word clouds can be a fun and powerful way to gather reactions and interpretations from a variety of people and make them part of your ensemble’s communication and honoring of the piece of music or text.

Leave a Reply

Your email address will not be published. Required fields are marked *