Week 10 Archive Research
Directions for sourcing and utilizing historical tweet data.
R-Shief Archive
One very useful source for historical social media data is the repository R-Shief, which has three data products:
- Twitter, Facebook, and Instagram data for the past 30 days
- Twitter data from 2010-2013 on hashtags relevant to global activist events
- Facebook, Youtube, and Website archives from 2008-2013
The second product is what interests me most as a potential corpus of material for annotatAR.
Data format
According to the website, R-Shief’s archive is accessible via their data visualization applications, where one can craft a chart and download the graphic to use.
They also have an incredibly informative table of relevant hashtags, with descriptions of what each abbreviation represents.
I have emailed R-Shief in the hopes that some of their archive will be available to use independently. Hopefully they will be able to provide an archive of data in JSON or CSV format that can be plugged in to the existing annotatAR architecture.
Right now, annotatAR translate incoming JSON data to a MongoDB database entry. In the context of utilizing historical data, the translatation step becomes unnecessary. However, translating the data archive to a database as a one-time operation could vastly improve the performance of the mobile site.
Aesthetic notes
There are many ways to interact with the more conventional object of historic record - museum exhibits, primary texts, historical re-enactments, photographs. Digital tools afford a powerful new layer of interactivity as media assets can be juxtaposed and arranged fluidly.
In relation to new media object such as the annotatAR augmented reality experience, digital media affords all of the physical interactions of the mobile web: swiping, tapping, scrolling, tilting – plus animations and UI changes that can be bound to these events.
There is something oddly compelling about before-and-after photos, where we see the ghosts of past photons pinned to the contemporary record. So also do the tweets form a figure of past actions and recent historicity against the ground of the familiar urban juice.
One direction I’ve considered for visualizing the tweets is rendering the video stream as letters, using the tweet text colored as if it were the backgroud video. Using the tweet text as a mask for the video would provide the pivotal visual metaphor of seeing the present through the lens of the virtual past.
Next steps
Test out the text masking idea using contemporary tweet data, and wait for answers from R-Shief.