Entity (2021) is an audio-visual installation generated from scraped tweets and carefully translated into sound and image.
The project began with a search of all tweets with the parameters:
From: ‘Anyone’
Location ‘Near me’
Language: ‘English’
Containing the exact phrase: ‘I feel’
Between the dates: ‘01/03/2020’ and ‘29/02/21’
The results were scraped and the text of the tweets along with the corresponding dates extracted, becoming a list of sentiments, sorted by the date they were sent into the world. Each one was read and assigned a category:
Strongly negative
Negative
Lightly negative
Neutral
Lightly positive
Positive
Strongly positive
Ambiguous
Mapping these results became a score. Putting them to paper and noting any clear feelings expressed along the way (anger, humour, fear etc).
Each day of the dataset is given around 5 seconds within the track, and each tweet made that day is represented by a MIDI note, it’s qualities corresponding to it’s assigned category. Secondary observations like ‘anger’, ‘humour’ are also represented by distinct textural noises. The velocity of each note is in accordance with the strength of the sentiment being expressed. Everything belonging to the ‘negative’ categories is panned somewhat left, and vice versa positive and right.
A third layer of abstraction comes with the video, as the dates tick by, surrounded by an animation of particles programmed to shift with the volume of the track.
Entity has been an exercise in documenting, categorising, and outputting pieces of semi-chaotic information into one continuous stream, mirroring the mechanism of social media itself. Individual units of content all competing for our attention but ultimately becoming flattened, compressed and forgotten in the ‘feed’.
It is an attempt to take the shifting, unpredictable, hum of activity that takes place online and make a small part of it more tangible.