[{"data":1,"prerenderedAt":12},["ShallowReactive",2],{"stack-tr-en-darts":3},{"tagline":4,"description":5,"strengths":6,"useCase":11},"The Python library that unifies time-series forecasting, from classic models to deep learning.","Darts, built by Unit8, exposes a single fit\u002Fpredict API across dozens of forecasting models: ARIMA, Exponential Smoothing, Prophet and Theta, but also lag-based regression (LightGBM, XGBoost) and deep networks such as N-BEATS, TFT, TiDE or TSMixer. Its TimeSeries object handles multivariate series, past and future covariates, probabilistic forecasting, backtesting and anomaly detection. It's our choice when several model families need a fair comparison on the same series before the best one goes to production.",[7,8,9,10],"Unified fit\u002Fpredict API: ARIMA, Prophet, boosting and deep learning are interchangeable.","Global models trained on thousands of series with past and future covariates.","Probabilistic forecasting: quantiles and intervals to quantify uncertainty.","Built-in backtesting and historical forecasts, plus an anomaly-detection module.","Sales and demand forecasting, inventory planning, energy consumption, capacity planning.",1792675979626]