New Update: A new robust LSTM model for M and X class flares is available! Note: It is not in the stable package yet, but you can download weights here.

SunFLAAR is an open-source Python package for comprehensive solar flare analysis, integrating deep learning and statistical methods. It provides automated tools for multivariate time-series forecasting, active region classification, and imbalanced data metrics to aid space weather research.

Please remember to acknowledge and cite the use of SunFLAAR in your publications.

Institutional Collaboration

SunFLAAR is in its foundational stages and is being adopted by researchers studying heliophysics and space weather phenomena. We welcome institutional collaborations to help integrate real-time observatory streams (like HMI/AIA). Contact us.

Latest Release

1.0.4
Check Release Notes
(2026-08-01)

Command line installation

$ pip install sunflaar

Report Bugs and Contribute

Handling imbalanced solar data (minor vs major flares) is challenging. If you encounter mistakes, errors, or bugs—or if you have a proposal for a new predictive architecture—please report it on the GitHub issue tracker. If you know how to fix a problem, we strongly encourage contributions.