IO

Mirroring

Creating and modifying living text in an infinite textual universe. Playing around with the concept of async communication, attention span, bots, and automated labyrinths. I've built a NodeJS bot, called Black Mirror, that is reading out phrases from the pages of an infinite library of n-dimensional books.

Tweets by Black Mirror Bot

Meanwhile, don't forget to listen to the mirrored wikipedia editing sequence of our governmental institutes behind the hexagonal corridors 

Changing rhetorics in an automated, bot-driven world: What if we keep only the loud parts in the tweets of the 44th & 45th presidents of the United States? Can we see difference and shifting rhetorics? What if we apply sentiment analysis on those redundant entities? (This goes to the future). A PHP script is used to access Twitter through OAuth, then some raw JS regexp tweaking applied on the texts.

Filtered tweets of US president Obama (first) & Trump (second). Only shouting (upper case) words and '!'s are left in the tweets. (right-click to see it in large)

So far, those tweets are filtered using the following expression:
tweet.replace(/\b(((https*\:\/\/)|www\.)[^\"\']+?)(([!?,.\)]+)?(\s|$))/g,' ').replace(/[a-z]/g,'_').replace(/[A-Z](?=_)/g,'_').replace(/[0-9]/g,'_').replace(/(U.S.)|(P.M.)|(A.M.)|( [A-Z] )/g, '').replace(/[^\w\s!]/g,'_').replace(/_{1,}/g,''); 

Oh my, who said regexp is not cryptic?