Summary

This repository serves as a comprehensive, community-driven collection of papers, code, and resources for Time Series Analysis (TSA). It targets researchers, engineers, and practitioners by organizing the latest advancements, particularly in Large Language Models (LLMs), Foundation Models, and Graph Neural Networks, offering a centralized hub for staying current in the rapidly evolving field of TSA.

How It Works

The project functions as a curated knowledge base, systematically categorizing and linking high-quality research papers and their associated code implementations. It covers core TSA tasks like forecasting, classification, imputation, and anomaly detection across diverse application domains. A key focus is on integrating cutting-edge techniques, such as LLMs and foundation models, with traditional and advanced methods like GNNs and Transformers, providing structured insights into their methodologies and applications.

Quick Start & Requirements

This repository is a curated list of research papers and code, not a deployable software project. Users should refer to individual linked papers and code repositories for specific installation instructions, dependencies (e.g., Python versions, libraries, hardware like GPUs), and setup procedures.

Highlighted Details

• Extensive Coverage: Features over 240 papers, with recent additions from major 2025-2026 conferences (ICLR, ICML, NeurIPS, KDD, AAAI, IJCAI).
• LLM Integration Focus: Deep dives into LLM applications in TSA, categorizing their roles as Inference Engines, Enhancers, or Hybrid Collaborators, and detailing various tokenization and prompting strategies.
• Diverse Applications: Covers critical domains including Finance, Healthcare, Energy, and Transportation.
• Foundation Models: Dedicated sections for emerging foundation models tailored for time series data.

Maintenance & Community

The repository shows signs of active maintenance, with a last update in April 2026. It relies on community contributions for content suggestions and updates. Specific community channels (like Discord/Slack) are not detailed.

Licensing & Compatibility

The project is licensed under the MIT License, which permits broad use, modification, and distribution, including for commercial purposes.

Limitations & Caveats

As a curated list, the repository's primary limitation is that it reflects the state of published research rather than providing a unified, tested framework. Users must evaluate the quality, applicability, and implementation details of individual linked projects. The rapid pace of research means content can quickly become dated despite regular updates.