Weather Premium
Weather Premium is a two-screen visual and sound installation that explores the links between weather modification programmes, AI simulation technologies and financial weather derivatives. The installation consists of two large screens showing different simulations, accompanied by a spatialised soundscape in the exhibition space.
The first screen presents a real-time climate prediction model with forecasts for three months into the future, showing parameters such as precipitation, cloud cover, pressure, wind and temperature - all layers that create a volumetric planetary atmosphere. The second screen shows a financial model of the globe, which translates weather data into economic value and identifies potential locations for weather intervention. This simulation uses the same pricing mechanisms that hedge funds use to calculate derivatives from historical weather patterns.
The sound composition integrates atmospheric phenomena such as raindrops, wind sounds and seasonal insects, which respond to live weather API data and financial disruptions displayed in visual simulations. By combining these natural soundscapes with data and forecasting systems, Weather Premium creates a tangible experience that reveals often forgotten phenomenological dimensions of the atmosphere.
A project by Luiza Crosman and Eduardo Castillo-Vinuesa Composer: Kat Austen Simulation Designer: Will Freudenheim Data Scientist: Artem Konevskikh Graphic design: Handi Kim Sound Engineer: Shinuk Kang
As a data scientist in Weather Premium project, I developed backend system that powers both visual simulations and sound installation. I created web service using FastAPI framework that continuously collects global weather data and forecasts, processes this information with NumPy to predict weather derivatives prices, and stores everything in PostgreSQL database. System then distributes these predictions to other parts of installation in real-time via API. Solution was built with focus on reliable data processing and efficient communication between different components of installation.
