# blue-yonder/tsfresh

Automatic extraction of relevant features from time series:

Repository: https://github.com/blue-yonder/tsfresh
Canonical: https://ross.abutalabs.com/products/tsfresh
Homepage: http://tsfresh.readthedocs.io
Language: Jupyter Notebook
License: MIT
License Family: permissive
Topics: data-science, feature-extraction, time-series
Last push: 2026-07-06T01:28:19+00:00

## Health v2 (maintenance only)
Score: 80/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 91, release rhythm 54, longevity 100
- inputs: {"age_days": 3598, "days_push": 59, "days_rel": 94, "gap_med": 234.0, "n_releases_24m": 3}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 9299, forks 1274 (observed 2026-08-28T04:10:30.432602+00:00)

## What it is
tsfresh is a Python package that automatically extracts hundreds of features from time series using algorithms from statistics, signal processing, and nonlinear dynamics. It includes a hypothesis-test-based filtering procedure to select only the features relevant for a given regression or classification task.

## Use cases
- extract features from time series for machine learning
- automate feature engineering for sensor data
- classify time series data
- forecast with time series features
- filter irrelevant features from time series
- build scikit-learn pipelines for time series

## When to choose
- you need many time series features computed automatically
- you want built-in relevance filtering via hypothesis tests
- you work in Python with pandas DataFrames and scikit-learn

## When to avoid
- you need deep-learning-based time series representations
- your data is not time series or event-sequence shaped
- you need extremely low-latency feature computation without parallelization setup

## Facets
- artifact type: library
- maturity: stable
- function: machine-learning, data-science, etl
- domain: data-science, machine-learning, time-series, analytics
- platform: python, cross-platform
- tags: time-series, feature-extraction, feature-selection, hypothesis-testing, scikit-learn

## Member repositories
- blue-yonder/tsfresh (main) score 80

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:10:30.432602+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-29T17:22:31.647111+00:00, confidence not recorded.
  - readme: https://github.com/blue-yonder/tsfresh (fetched 2026-08-28T04:10:30.432602+00:00, sha 43d7521d8dd8)
  - homepage: http://tsfresh.readthedocs.io (fetched 2026-08-29T08:22:35.688873+00:00, sha 5db8b31f0370)
  - registry_pypi: https://pypi.org/pypi/tsfresh/json (fetched 2026-08-29T08:22:35.699010+00:00, sha 123d34a0702c)
- Data as of 2026-08-30T08:39:29.467469+00:00.