Apresentador(a): Italo Pinto Rodrigues

Título do Trabalho: Experience Report on the Prototyping of a Mobile System for Geomagnetic Storm Forecasting: Challenge-Based Learning

Resumo: This study, conducted as part of the NASA Space Apps Challenge, exemplifies Challenge-Based Learning, allowing students to develop innovation skills while addressing the significant risks of geomagnetic storms to global electronic infrastructures, which have potential economic impacts of $2.6 trillion. Utilizing a Long Short-Term Memory (LSTM) neural network, we analyzed data from the DSCOVR satellite, navigating its limitations. Our methodology entailed developing a Solar Classification (SC) index from the Z component of the mag-
netic field (Bz) to address data inconsistencies. We trained the LSTM with 80% of the refined

dataset and validated it with the remaining 20%, achieving a Root Mean Square Error (RMSE)
of 0.8724%. This research, arising from a collaborative and competitive educational setting,
highlights the effectiveness of team-based approaches in tackling complex scientific challenges
and demonstrates the potential of AI in improving space weather forecasting and enhancing
public preparedness readiness.