Week 5 Invisible Monument

Outcomes from deployement at Invisible Monument launch event in Boston.

Lead up to deployment

In the past week, I had a development push to meet the development goals that were most important to be able to demonstrate annotatAR’s utility as a docmentation tool. I met those goals for the most part. I was able to implement:

• Typography and color changes for better legibilty
• Stripping the tweet text of URLs
• Screen capture button to save an image of the canvas element
• Panning with accelerometer data - incomplete implementation
• Integrating stream and REST Twitter API data

Ideally, I want to use accelerometer data from the device to allow the end user to pan around - to change the position of the tweets to create the illusion that they are positioned in three dimensional space. While I wasn’t able to achieve a reliable 3D effect, I started to experiment with the ondevicemotion JavaScript event, and accellerationIncludingGravity on the X and Y axes. While I was able to achieve somewhat fluid UX on my laptop, the calibration did not translate properly to my Galaxy 4G device.

Invisible Monument

The Invisible Monument launch event, #LuncheonOnTheGrass took place on 30 Sept 2015, in Dewey Square, Boston. Invisible Monument is a locative sound installation created by Halsey Burgund and Lara Baladi, that cn be experienced using an iOS application. Various sond elements are triggered by geolocations around Dewey Square.

I met the team on site for the launch event and started taking screen captures with annotatAR. I used three different Twitter tags: #occupy, #the99, and #luncheononthegrass. I adjusted the “die-off” rate for each of the tweets according to how often the tag was used. As the tweets get older, they become smaller and more transparent.

Historic tweets

I would like to be able to overlay historic tweet data as well as current tweets. However, the Twitter REST API allows access to tweets created within the past couple of weeks only.

Third-party services such as GNIP are social media brokers from which you can purchase historic data. The base rate for historic tweet data starts at $1.1k and increased depending on length of time and number of tweets returned by the query.

While this is cost-prohibitive for the time being, it does add a layer of utility to annotatAR. I have already implemented the ability to download the database of tweets as a CSV file, which could be quite useful in capturing an independent record of a hashtag.

Additionally, I spoke to one of the participants in #LuncheonOnTheGrass who directed me to an archive of tweets from the Occupy movement, available at Occupy Research and R-Shief. These might be a useful corpus to draw on in future iterations.

Future of Invisible Monument

Lara and team were quite excited by the output from annotatAR, and will use my documentation of #LuncheonOnTheGrass as part of the ongoing archive for Invisible Monument / Vox Populi.

We began to discuss how annotatAR might be integrated with Halsey’s Roundware platform, to provide an integrated experience with the audioscape. Future directions for Invisible Monument are quite exciting, and I hope to continue to work with them. However, I think this will only be possible of Halsey ports Roundware to Android or the mobile web.

Near term goals

More immediate goals for annotatAR include:

• Bug fixing - hashtags on the stream seemed not to work on the website (though the stream worked on the local version)
• Bug fixing - houston doesn’t seem to work, maybe due to special characters?
• Fluid accelerometer use
• URL routing for hashtag display
• Automation or parameterization of annotatAR UX elements (ie age til die-off)
• Capturing additional tweet data (ie user, associated photos) for posterity

Once these goals are achieved it may be time to turn to text analysis and a desktop site.