Three trillion location records reveal how 500,000 simulated people move through San Francisco

Chanuka Algama, Taylor Anderson, Henrique Ferraz de Arruda, Andrew Crooks, Nathan Holt, Erfan Hosseini Sereshgi, John Hunter, Hamdi Kavak, Lance Kennedy, Yueyang Liu, Dieter Pfoser, Sandro Martinelli Reia, Doug Taylor, Mauryan Uppalapati, Boyu Wang, Carola Wenk, Andreas Züfle

SF-LIFE simulates 500,000 agents moving through the San Francisco Bay Area for 70 days, generating 3 trillion location records at 1-second resolution. The dataset captures realistic daily schedules (shopping, work, socializing) mapped onto real transit networks—buses, trains, bikes, cars, walking—using actual GTFS data from 40+ agencies. Unlike real movement data, it's complete, noise-free, and ethically sourced, enabling researchers to test transportation models and urban planning ideas without privacy concerns.