TIMPy

TIMPy is an open-source Python library for time-aware incremental learning and time-series analysis. It provides tools for online model training, streaming data processing, and time-centered evaluation, enabling real-time analytics in research and industry. TIMPy emphasizes a lightweight, extensible API designed to resemble familiar scikit-learn patterns while supporting continuous updates to models as new data arrive.

TIMPy originated as a community project in the early 2020s and reached a stable 1.0 release by the mid-2020s. The project is hosted on a public code repository with active contributors from academia and industry. Documentation and tutorials accompany the releases, guiding users through installation, data preparation, model selection, and evaluation on streaming datasets.

Key features include online and incremental learners for classification and regression, support for streaming data sources, and time-aware evaluation methods such as rolling and expanding window cross-validation. The library offers data preprocessing and feature extraction utilities suitable for time-series data, integration with common Python data stacks (pandas, NumPy), and a pipeline interface for building end-to-end workflows. Additional capabilities cover model persistence, experiment tracking, and basic visualization of streaming metrics.

TIMPy aims to fill gaps in traditional batch-oriented toolkits by providing robust handling of concept drift, real-time scoring, and incremental persistence. It remains under active development, with ongoing work focusing on performance improvements, broadened algorithm coverage, and expanded documentation.