Los Angeles Quake Prediction App of QuakeCode harnesses advanced artificial intelligence (AI) and machine learning (ML) to provide accurate earthquake risk forecasts for Los Angeles and nearby regions. Los Angeles, with nearly 19 million residents in its metropolitan area, lies along the San Andreas Fault System, one of the world’s most active and complex seismic zones. While Northern California was devastated by the 1906 San Francisco earthquake, Southern California remains on the verge of a major seismic event, having not experienced a rupture of similar magnitude in over a century. This heightened risk makes timely and reliable forecasts essential to protect lives, infrastructure, and improve disaster preparedness and response.

Our model combines supervised deep learning and gradient boosting using XGBoost. Trained on over 110,770 earthquakes (Mw ≥ 2) recorded from 1990 to 2025 in Southern California, it analyzes seismic data—such as earthquake occurrence time, magnitude, depth, seismic energy, and b-value—using a sliding window technique. This approach allows the model to capture evolving seismic activity patterns over time. From this data, sixteen AI-derived features are extracted to predict the maximum expected earthquake magnitude for the next 60 days. The XGBoost model, optimized with 3000 estimators, learning rate 0.03, and max depth 6, achieves an R² of 0.98, demonstrating extremely high predictive accuracy and reliability.

QuakeCode also offers interactive 2D, 3D, and 4D visualizations of seismic patterns and forecasts using powerful Python libraries such as Matplotlib, Plotly, Geopandas, and mpl_toolkits.mplot3d. These visualizations enable researchers, engineers, government agencies, and the public to better understand complex seismic trends, spatial distributions, and prediction results, facilitating informed decision-making and risk assessment.

The app is built with a robust technology stack, including Python, TensorFlow, Keras, pandas, and XGBoost. Development and experimentation are conducted in flexible and scalable environments such as Jupyter Notebook, Google Colab, Anaconda, and Xcode. This infrastructure supports continuous improvement, ease of collaboration, and deployment in diverse settings, making it accessible for researchers and practitioners worldwide.

Beyond earthquake prediction, QuakeCode contributes to environmental sustainability by supporting the monitoring of underground CO₂ storage sites. Carbon capture and storage (CCS) is a crucial strategy for mitigating climate change by reducing atmospheric greenhouse gases. Seismic monitoring ensures the integrity and safety of these storage sites over time, helping to detect potential leaks or instability early. QuakeCode’s seismic analysis tools thus provide valuable insights for both hazard mitigation and environmental protection.

QuakeCode’s mission is to reduce earthquake hazards, protect lives, and strengthen infrastructure resilience. By delivering data-driven, AI-powered forecasts and visualizations, the app empowers policymakers, scientists, engineers, emergency responders, and communities to make better-informed decisions that promote safer, more prepared, and sustainable societies both locally and globally.

QuakeCode is intended for educational, research, and public awareness purposes only. It does not collect, store, or share any personal data, ensuring user privacy. Furthermore, while the app’s forecasts are scientifically grounded and highly accurate, they should not be considered official emergency warnings. Users should always rely on guidance from local authorities and authorized seismic monitoring organizations in the event of earthquake threats.