Open Software

All software produced by my DMLab is publicly available and free to use with appropriate citation. Our primary code repository is hosted on Bitbucket at bitbucket.org/gsudmlab.

Machine Learning Toolkits

MVTS-Data Toolkit

A domain-independent Python package for preprocessing multi-class multivariate time series (MVTS) data. Provides 48 statistical feature extractors, automated parallel feature extraction, normalization, imputation, and sampling methods for class-imbalanced datasets. Installable directly via pip.

• 📄 Paper: Ahmadzadeh, Sinha, Aydin & Angryk, SoftwareX, 2020
• 💻 Bitbucket: bitbucket.org/gsudmlab/mvtsdata_toolkit
• 💻 GitHub Mirror: github.com/AzimAhmadzadeh/mvtsdata_toolkit
• 📦 PyPI: pip install mvtsdatatoolkit — pypi.org/project/mvtsdatatoolkit

Solar Physics & Space Weather Tools

SDO Image Parameters API

A public API providing access to a large image parameter dataset extracted from NASA SDO/AIA instrument, covering January 2011 onward at 6-minute cadence across nine wavelength channels. Supports content-based image retrieval and solar event tracking tasks.

• 📄 Paper: Ahmadzadeh, Kempton & Angryk, ApJS, 2019
• 💻 Bitbucket: bitbucket.org/gsudmlab/imageparams_api

TaMag — Magnetogram Transformation Module

A software module for transforming and preprocessing solar magnetogram data from NASA SDO/HMI, developed to support machine learning pipelines for space weather forecasting.

• 💻 Bitbucket: bitbucket.org/gsudmlab/tamag

Spatiotemporal Interpolation Tools

Software implementing spatiotemporal interpolation methods for filling gaps in solar event trajectories, supporting data quality improvement in solar physics analytics pipelines.

• 📄 Paper: Filali Boubrahimi et al., ApJS, 2018
• 💻 Bitbucket: bitbucket.org/gsudmlab/stinterpolation

Flare Prediction Pipeline

Source code for solar flare prediction workflows developed in the DMLab, including data preprocessing, model training, and evaluation pipelines built on the SWAN-SF benchmark dataset.

• 💻 Bitbucket: bitbucket.org/gsudmlab

Feature Subset Selection on SWAN-SF

Code for systematic feature selection experiments on the SWAN-SF heliophysics time series dataset, evaluating 24 feature selection methods for solar flare forecasting.

• 💻 Bitbucket: bitbucket.org/gsudmlab/featuresubsetselection_on_swan-sf

DMLab Core Library

DMLab Library

Core utility library supporting DMLab research projects across data mining, spatiotemporal analysis, and machine learning pipelines.

• 💻 Bitbucket: bitbucket.org/gsudmlab/dmlablib

All Repositories

The complete list of DMLab open-source repositories is available at:
bitbucket.org/gsudmlab

For questions about any software or collaboration inquiries, please contact us.