# uber/orbit

A Python package for Bayesian forecasting with object-oriented design and probabilistic models under the hood.

Repository: https://github.com/uber/orbit
Canonical: https://ross.abutalabs.com/products/uber-orbit
Homepage: https://orbit-ml.readthedocs.io/en/stable/
Language: Python
License: NOASSERTION
License Family: other
Topics: python, forecasting, bayesian, exponential-smoothing, pyro, stan, pystan, probabilistic-programming, probabilistic, forecast, orbit, time-series, regression, arima, changepoint, bayesian-methods, bayesian-statistics, machine-learning, regression-models, pytorch
Last push: 2026-05-22T22:21:47+00:00

## Health v2 (maintenance only)
Score: 83/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 83, release rhythm 73, longevity 100
- inputs: {"age_days": 2430, "days_push": 103, "days_rel": 103, "gap_med": 80, "n_releases_24m": 2}
- flags: no_license
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 2069, forks 148 (observed 2026-08-28T04:06:10.233299+00:00)

## What it is
Orbit is a Python package for Bayesian time series forecasting and inference with an initialize-fit-predict interface. It uses probabilistic programming backends like Pyro and Stan (via cmdstanpy) under the hood.

## Use cases
- forecast time series with bayesian methods
- fit exponential smoothing models in python
- predict future sales or demand with uncertainty estimates
- detect changepoints in time series data
- run bayesian regression on temporal data
- compare forecasting models like arima and ets

## When to choose
- you need probabilistic forecasts with uncertainty quantification
- you want a scikit-learn-like fit/predict API for time series
- you need bayesian structural time series or exponential smoothing models

## When to avoid
- you need lightweight classical forecasting without bayesian sampling overhead
- you cannot install cmdstanpy or its Stan toolchain dependencies
- you need real-time low-latency predictions rather than batch forecasting

## Facets
- artifact type: library
- maturity: stable
- function: machine-learning, data-science, math
- domain: time-series, machine-learning, data-science
- platform: python
- tags: bayesian-forecasting, time-series, probabilistic-programming, pyro, stan, exponential-smoothing, changepoint-detection

## Member repositories
- uber/orbit (main) score 83

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:06:10.233299+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 2026-08-30T02:56:46.455211+00:00, confidence not recorded.
  - readme: https://github.com/uber/orbit (fetched 2026-08-28T04:06:10.233299+00:00, sha d71b9ecfdaee)
- Data as of 2026-08-30T08:39:29.467469+00:00.