#nightTweets
Sarah Hallacher
#nightTweets uses your sleep data to power your social media outreach on twitter, so you'll never have an unproductive night's sleep again!!!
http://sarahmak.es/
#nightTweets is an exploration of productivity under the satirical assumption that we are 'not productive enough' during the day. Using EEG data, #nightTweets sends messages from the unconscious mind via Twitter. Each stage of sleep (light, deep and REM) corresponds to a different type of twitter conversation; including tweeting at other users, discussing trending topics, retweeting, sending direct messages, and use of hashtags. By outsourcing a daytime task to our sleeping brain, #nightTweets invades the private and life-sustaining act of sleep; and disrupts a time frame during which we're not meant to be productive.
Background
I did a great deal of research about what happens to our brains and bodies during different stages of sleep.
Audience
My target audience is my group of followers on twitter. I've run the program live twice and had a series of reactions from my followers and from twitter strangers: 19 retweets, 18 tweets favorited by others, 24 replies to my tweets, 6 direct message responses, 6 followers lost, and 4 followers gained.
User Scenario
#nightTweets runs in Processing using the Twitter4J & Rita libraries, and uses data from the Zeo sleep tracker. The program analyzes my twitter account to create a "vocabulary" to use when it creates new tweets or messages, as well as gain a knowledge of my followers. Every five minutes, the program receives data from the Zeo and a different action is taken on twitter depending on which stage of sleep I'm experiencing.
Implementation
In light sleep, the program finds the top 10 local trending topics, chooses one hashtag, and generates a tweet using that hashtag. In REM sleep, the program uses my keystroke log data (collected using Backtrack) and tweets random selections. In deep sleep, the program chooses one of my followers at random and generates a direct message to send to them.
I also programmed tweets to announce when I fall asleep, when I wake up in the middle of the night, if I can't fall back to sleep, and when I wake up for the day. My profile photo, header image and profile colors change to a nighttime design when I fall asleep and back to my normal daytime design when I wake up.
Conclusion
A big challenge was knowing that I was going to irritate people. I'm not much of an over-sharer on social media and it concerned me that my change in behavior would not be welcome. In the future, I'd like to build out my program to take advantage of all the features the Twitter4J library offers. I'd also be interested in setting up the program for another user so we can talk to each other via twitter in our sleep.
Background
I did a great deal of research about what happens to our brains and bodies during different stages of sleep.
Audience
My target audience is my group of followers on twitter. I've run the program live twice and had a series of reactions from my followers and from twitter strangers: 19 retweets, 18 tweets favorited by others, 24 replies to my tweets, 6 direct message responses, 6 followers lost, and 4 followers gained.
User Scenario
#nightTweets runs in Processing using the Twitter4J & Rita libraries, and uses data from the Zeo sleep tracker. The program analyzes my twitter account to create a "vocabulary" to use when it creates new tweets or messages, as well as gain a knowledge of my followers. Every five minutes, the program receives data from the Zeo and a different action is taken on twitter depending on which stage of sleep I'm experiencing.
Implementation
In light sleep, the program finds the top 10 local trending topics, chooses one hashtag, and generates a tweet using that hashtag. In REM sleep, the program uses my keystroke log data (collected using Backtrack) and tweets random selections. In deep sleep, the program chooses one of my followers at random and generates a direct message to send to them.
I also programmed tweets to announce when I fall asleep, when I wake up in the middle of the night, if I can't fall back to sleep, and when I wake up for the day. My profile photo, header image and profile colors change to a nighttime design when I fall asleep and back to my normal daytime design when I wake up.
Conclusion
A big challenge was knowing that I was going to irritate people. I'm not much of an over-sharer on social media and it concerned me that my change in behavior would not be welcome. In the future, I'd like to build out my program to take advantage of all the features the Twitter4J library offers. I'd also be interested in setting up the program for another user so we can talk to each other via twitter in our sleep.