STUMPY
A Powerful and Scalable Python Library for Modern Time Series Analysis
The Whole is Greater than the Sum of Its Parts
STUMPY is a powerful and scalable Python library for modern time series analysis and, at its core, efficiently computes something called a matrix profile. The goal of this multi-part series is to explain what the matrix profile is and how you can start leveraging STUMPY for all of your modern time series data mining tasks!
Note: These tutorials were originally featured in the STUMPY documentation.
Part 1: The Matrix Profile
Part 2: STUMPY Basics
Part 3: Time Series Chains
Part 4: Semantic Segmentation
Part 5: Fast Approximate Matrix Profiles with STUMPY
Part 6: Matrix Profiles for Streaming Time Series Data
Part 7: Fast Pattern Searching with STUMPY
Part 8: AB-Joins with STUMPY
Part 9: Time Series Consensus Motifs
Part 10: Discovering Multidimensional Time Series Motifs
Part 11: User-Guided Motif Search
Part 12: Matrix Profiles for Machine Learning
The Origin Story
STUMPY derives its name from its algorithmic predecessors (i.e., STAMP and STOMP) and pays homage to other foundational Python numerical computing packages (i.e., NumPy, SciPy, and Numba). Originally, STUMPY was conceived in Ann Arbor, MI, which is often known as the “City of Trees”, and so a name was born.
Resources
UCR Matrix Profile
STUMPY Matrix Profile Documentation
STUMPY Matrix Profile Github Code Repository
