Continuing on the previous post, Tweetscapes takes on Twitter as well. However, there are notable differences. While Twinthesis can be used as an instrument and offers a few interaction possibilities, Tweetscapes is meant to be listened to, experienced without any direct possible interaction. However, the tweetscapes soundscape has been used as a base for improvisation in live music.
Tweetscapes sonifies hashtags, the keywords that are used on Twitter to tag messages. The project only sonifies Tweets originating in Germany and thus, to be sonified, the Tweets have to adhere to at least one of two conditions:
- The tweet is geotagged to a location in Germany, or
- The tweet contains at least one word in German and the app language is set to German.
The location of Twitter hashtags controls the stereo position: the more to the west the tweet comes from, the more to the left the associated sound is played. The hashtags values are summed and further mathematical operations are applied, which results in a unique number for each topic. This number (between 1 and 1.000.000) controls the parameters of the sound generation, making each sound unique and surprising. The ratio of vowels in a hashtag determines the softness/harshness of the sound: the more vowels, the more harmonious. When the same hashtag is trending, e continuous background noise is created to make this trend audible. @Replies, answers directed to another specific user, are sonified using whispersounds. The authors say that during the night there are more @replies than during the day. Tweets without hashtags or without a direct recipient are sonified with short generic sounds. The number of followers of the tweeter determines the volume: the more followers, the louder the sound. When a tweet is retweeted, the sound is echoed. When a message is popular, this results in the sound being denser and the fade out longer.
In the unusual event that Twitter is down, tweetscapes plays whalesounds. This refers to the image Twitter shows when it is down.
Florian Eitel developed a custom tool inRuby to query the Twitter API. The audio part was dealt with by Anselm Venezian Nehls & Dr. Thomas Hermann using SuperCollider.
Besides sonifying the Twitter activity, there is also a visual component, handled through Versum by Tarik Barri, based on Max/MSP, Java, and “a number of secret ingredients”. the whole streaming process is done via Flash Media Encoder.
You can find more information about the project here. The project’s funding has now finished so it is not online anymore. You can still listen and see the video below to give ou an idea.
One Reply to “Tweetscapes: sonifying Germany’s twitterscape”