deep learning algorithms to reconfigure sacral tectonics ​ aiming to find alternative ways to visualize the rapid change of public discourse on digital platforms, the project uses architectural imagery to showcase user opinions on twitter regarding religious tendencies.   the final work, adapting in realtime and accessible via web, uses various collaborating stylegan neural_nets to visualize an altar of technocarcy. the work dynamically superimposes characteristics of different religions, origins and timeperiods depending on data accessed by twitter api

through analysis of 9 000 sacral p ojects from 700BC to 2019AD, the altar_neuralnet generates new interpretations of existing religious  architectures. the architecture of all major religions was encapsulated through the training of seperate stylegan2ada neuralnets. depending on the distribution of keywords within several religious debates on twitter, the different deep learning algorithms were mixed with according weighting and used to generate the realtime 3d output via a unity interface.

depending on time of the day and political events, the final work changed appearance radically. the project was public from 1.3.2021-1.5.2021 and became the start of several research projects dealing with efficient translation of 2d data in