- Plotly: Interactive time series charts.time series
Forecasting & Modeling
- statsmodels: ARIMA, SARIMA, state-space models, Statistical Tests for stationarity and autocorrelation.
- pmdarima: Auto-ARIMA model selection and diagnostics.
- prophet (by Meta): Decomposable additive models, handles trend, seasonality, and holidays easily.
- ruptures: Change point detection for time series.
Feature Engineering & Preprocessing
- tsfresh: Automatic Feature Extraction from time series for machine learning.
- featuretools: Feature engineering (can be used with time series if structured correctly).
- Scikit-Learn
- PyOD: Comprehensive library for outlier and anomaly detection (can be applied to time series).
Others:
- Darts