Racks (2026) is an interactive installation documenting abandoned and dismantled bicycles across four London boroughs. The project examines bicycles that once belonged to someone, often for years, before being stripped for parts through theft and left attached to public racks throughout the city. Many still retain traces of former ownership, including stickers, scratched paint, custom modifications, locks, and signs of prolonged use. Although physically present within the urban environment, their original purpose and relationship to ownership have collapsed.
The project was initially developed in response to the impact statements published by Stolen Ride, where victims of bicycle theft describe the emotional and practical consequences of losing their bicycles. These testimonies became the conceptual foundation of the work. I later travelled across London to locate and document dismantled bicycles in the same boroughs from which these statements originated, integrating the testimonies into the archive according to location.
The bicycles exist in a suspended state between use and abandonment.
Because they are technically still private property, councils are often unable to remove them, despite their continued deterioration in public space. Over time, they become fragmented urban artefacts embedded within the city’s infrastructure.
Technically, the work combines photogrammetry, 3D scanning, interactive mapping, moving image, and web-based installation. Each bicycle was digitally archived and animated using Blender, AI-assisted workflows, and custom-coded interactive systems. Alongside the scans, moving image sequences capture the bicycles’ ongoing “agony” within London’s urban environment, positioning them as overlooked witnesses to systems of theft, neglect, ownership, and decay.
Racks was developed as a browser-based digital archive and interactive installation combining photogrammetry, 3D animation, web technologies, moving image, and spatial sound. The project documents abandoned and dismantled bicycles across four London boroughs through a workflow that merged physical field research with digitally generated environments and interactive systems. The bicycles were first located and documented on-site across London using photography, video recording, and photogrammetry scanning methods. Multiple image sets of each bicycle were captured and processed into 3D assets using Polycam and photogrammetry workflows. The resulting scans were cleaned, optimised, and reconstructed inside Blender, where geometry, textures, lighting, and materials were refined for real-time web presentation.
Additional modelling adjustments were made manually to preserve damaged structures, fragmented components, and traces of ownership such as scratches, locks, and missing parts. Animation systems for the bicycles were developed in Blender using physics-based movement simulations and custom Python-assisted workflows. Individual bicycle components such as wheels, pedals, chains, cranks, and frames were rigged and animated independently to create unstable mechanical movement responding to generated parameters. These motion systems were designed to simulate instability, deterioration, and mechanical stress, creating the sense that the bicycles remained in an ongoing state of collapse. Rendering tests, environmental lighting, fog systems, and camera choreography were also developed within Blender before export.
The interactive archive itself was built as a web-based installation using JavaScript and Three.js. Custom-coded systems were developed for loading and displaying GLTF 3D models in real time within the browser. The project utilised OrbitControls, dynamic lighting systems, environmental fog, shadow mapping, proximity-based interactions, and responsive camera systems to allow viewers to navigate through the digital archive spatially. HTML, CSS, and JavaScript were used to construct the interface, archival layouts, borough-based navigation systems, and embedded impact statements connected to individual bicycle locations. Sound layers were incorporated to reinforce the mechanical and emotional atmosphere of the archive.
AI-assisted workflows and ChatGPT-supported coding processes were used throughout development, particularly during scripting, interactive prototyping, motion systems, debugging, optimisation, and the construction of generative behaviours across the web environment and animation pipeline.
Project Page | Timothy Yufit | Instagram
