Skip to content
Get packs

Prophet

Open-source forecasting tool for time series with multiple seasonal patterns and trend changes.

GitHub Download v1.5.0
Stars
20.5k
Forks
4.6k
Licence
MIT
Language
Python
Last update
15 days ago
Type
Library

What it is

Prophet is a forecasting procedure for time series data, based on an additive model that fits non-linear trends along with yearly, weekly, and daily seasonality, plus holiday effects. The README says it works best with data that has strong seasonal effects and several seasons of history, and that it is robust to missing data and shifts in trend. It is released by Facebook's Core Data Science team and is available for R (CRAN) and Python (PyPI).

  • Forecast a daily or weekly business metric that has yearly and weekly patterns
  • Add holiday effects to a forecast using the built-in holiday support
  • Run cross-validation and performance metrics on a forecast (cross_validate and performance_metrics are mentioned in the changelog)

Great for

  • Analysts forecasting business metrics with weekly or yearly seasonality
  • Python or R users who want a forecasting model they can run without deep statistics expertise
  • Teams with several seasons of historical data and known holiday effects

Not ideal for

  • Users who need the maintained feature set to grow: the README says v1.4.0 entered maintenance mode, with only bug fixes, dependency bumps, and R/Python parity changes planned
  • Very short histories or series without clear seasonal patterns, since the README says it works best with strong seasonality and several seasons of data
  • Users who need a Spark/JVM implementation: that is Helter Skelter, a separate simplified reimplementation by Amadeus, not this project
Step-by-step guide: how to use ProphetInstall, first run, examples, tips and FAQs

How to get it

  1. 1

    Install with a package manager

    You'll need: Python 3.7 or newer (per README, as of v1.1) · pip or conda

    Install from PyPI

    python -m pip install prophet

    Install from conda-forge

    conda install -c conda-forge prophet
  2. 2

    Run it

    • Install the Python package with pip or conda
    • Open the quick start guide linked in the README: https://facebook.github.io/prophet/docs/quick_start.html#python-api
    • Prepare your data with a date column and a value column as the quick start describes, then fit and forecast
  3. 3

    Get the most out of it

    • Add holiday effects to improve forecasts where known events matter (holiday support is part of the model per the README)
    • Use cross_validate and performance_metrics to evaluate forecast accuracy (changelog references both)
    • Use the cmdstanr backend in R if you want an alternative Stan backend (documented as experimental in the README)
    • Check the release notes for bug fixes before upgrading, since v1.5.0 changed where performance_metrics lives (now prophet.metrics, still importable from prophet.diagnostics)

Downloads

Source code (.zip)main branch · straight from GitHubRelease v1.5.0Release notes and sourceOfficial website & docsfacebook.github.io/prophet

Strengths

  • Handles multiple seasonality patterns and non-linear trends in one additive model
  • Robust to missing data and trend shifts, according to the README
  • Available through both pip and conda-forge for Python, and CRAN for R

Trade-offs

  • The README states the project is in maintenance mode as of v1.4.0, with no new features planned
  • The CRAN version is described as fairly outdated; the README recommends installing from GitHub for the latest release
  • Building from source needs a C/C++ toolchain and, on Linux, at least 4GB of memory during install per the README

How it compares

ProjectStarsTypeLevelUpdatedBest for
prophet (this)20.5kLibrarybeginner15 days agoAnalysts forecasting business metrics with weekly or yearly seasonality
LightGBM18.9kLibraryintermediate4 days agoData scientists working with tabular data who want a fast gradient boosting tool
ML-For-Beginners91.9kLearning resourcebeginner5 days agoBeginners who want a structured, self-paced introduction to machine learning
ds-cheatsheets16.4kCurated listbeginner2.3 years agoBeginners who want quick reference sheets for Python or R

Stats synced from GitHub. The assessment above was drafted with AI from the project's README on 2026-10-29 and may miss recent changes — check the README before relying on a command. MIT licence: free to use, modify and distribute, including commercially, as long as you keep the copyright notice.