Summary of project:
A microphone installed in the DFI lab records volume levels through Processing. Processing then averages the volume levels over a 10-minute time frame. Every 10 minutes a Twitter account linked to the Processing sketch will tweet the current volume level (currently, we have 4 levels). If there has been no change in volume since the previous 10 minutes, there will be no new tweet.
We plan to use a tablet to run our sketch, and to leave that tablet somewhere secure in the DFI room (likely a locker). The tablet will continuously run the sketch as an application, providing us with more data in the form of tweets. We need to ensure the tablet is always plugged in, which could prove difficult if we put the tablet in a locker, but we will test that out.
First few versions of our code:
a. This version worked, but there was no function in the code to only send out a tweet every 10 minutes. This meant that once we ran the sketch, it tweeted the raw input data, every value read from the microphone. This meant that the sketch tweeted rapidly and often tweeted the same thing, pushing over the daily limit of 1,000 tweets. This caused Twitter to reject our tweets and essentially suspend our account for a few hours. This made testing a challenge.
- First iteration: https://github.com/wrightlauraa/repo1/blob/master/Twitter%2BProcessing
b. This version is similar, but with functions built-in so that our sketch only sent tweets every 10 minutes. This meant that we were no longer overloading Twitter.
It also has volume averaging, which calculates the volume over a period of 10 minutes to ensure that outliers don’t skew the data. This volume average determines which tweet to send out (there are currently 4 to choose from: quiet, loud, really loud, super loud). We will likely add more than 4 later and make them more specific.
- Second iteration: https://github.com/km13oj/project_3/blob/master/working_average_w_twitter
Outstanding problems to address:
a. We need to figure out how to turn a Twitter RSS feed of our account (@VolumeBot2013) into an XML file.
b. We need to figure out how to turn an XML feed into a data visualization with Processing.