The installation’s development began with an analysis of data connected to the city of Milan. In particular, it focused on a layer of information available on Twitter, in a way similar to what was explored in Amygdala.

In each cycle, tweets published in the city of Milan are analyzed and assigned an emotional value - an index of the positivity or negativity of the content - obtained through semantic text analysis. Their geolocation is also retrieved. The analyzed tweets are then clustered and displayed on the generated map based on their geolocated position, with colors reflecting their emotional impact.

Simultaneously, GDELT, a service that monitors news published by media outlets, news agencies, and newspapers around the world in real time, is used to filter news coming from Italy and store the emotional value assigned to them. The average of these values becomes a sort of “background noise” that deforms and modifies certain aspects of the generation of the visual and sonic elements.

At the heart of the installation lies a simulation system that generates a continuously evolving map of the city. Its design uses a series of algorithms inspired by emergent and self-organizing structures found in certain biological systems. In particular, ants, termites, and some types of fungi build extremely complex structures through simple communication processes. Our idea drew on some of these strategies to create a network-based structure for generating a continuously transforming map of Milan, starting from the real city layout but with the freedom to change and adapt according to incoming impulses—potentially in an endless process.

The sound of Falin Mynd is based on the idea of musical cryptography and the contrapuntal compositional technique of the canon. Sounds intertwine melodically and rhythmically, progressively imitating one another to create polyphonic flows in constant transformation.

Musical cryptography is a sequence of notes encoding non-musical content, typically a text. This technique has been used by various composers to embed their names or initials into compositions.

A canon is a contrapuntal composition that layers one or more imitations over an initial melody. The initiating voice is called the “antecedent,” while the following voices are the “consequent.”

All design and programming were entirely developed using Max/MSP and JavaScript, later integrated into Ableton Live via M4L. The soundscape is generated in real time using Twitter texts and analyzing their “positive–negative” emotional component through the GDELT project.

Technically, the texts of tweets, converted into binary format, produce numerical sequences used as rhythmic triggers for musical notes—also derived from text conversion. These melodic structures follow one another through progressive imitation until a new tweet generates a “new melodic pattern.”

The notes extracted from tweets are derived from extended chords beyond the triad form, following harmonic progressions over time. Positive and negative values from GDELT are associated with a major or minor scale to express an overall “global” emotional atmosphere, leveraging the expressive qualities of the scales’ interval structures.