Earthquakes have historically inflicted significant losses, with the period from 2000 to 2019 witnessing approximately 721,000 fatalities and US$636 billion in economic damages worldwide. Catastrophic earthquakes can wreak havoc on multiple structures and infrastructure systems simultaneously, leading to profound economic repercussions. While some losses can be mitigated through enhanced construction practices or transferred via the (re)insurance market, rapid response in certain scenarios is critical. Parametric insurance emerges as a strategic tool in such instances, designed to provide predetermined payouts for specific earthquake scenarios, often utilizing publicly available data for near-real-time responsiveness. However, parametric solutions are susceptible to basis risk, prompting efforts to minimize this discrepancy through evolutionary computation, machine learning, and biased-randomized algorithms. This study aims to help insurers by predicting losses produced by earthquake in Chile rapidly and with minimal uncertainty. Through a thin plate spline (TPS) advanced interpolation method this research seeks to advance understanding and address key challenges in parametric insurance design and implementation.