Prophet — Time Series Forecasting by Facebook (Meta)#
An open-source library for decomposable time-series forecasting with trend and seasonality.
Important
✨ AI-generated content. This page was written with the assistance of an AI language model and is provided as a learning aid. Despite careful review, it may still contain mistakes, omissions, or out-of-date information. Whether you are new to the topic, a team lead, or a senior practitioner, treat it as a starting point rather than an authoritative reference: read it critically and independently verify anything you act on (code, commands, figures, and factual claims) against official documentation and primary sources before relying on it.
What it is#
Prophet is an open-source forecasting library from Facebook (now Meta) that makes time-series forecasting simple, scalable and interpretable — designed for business data with trends, seasonality and holidays, and usable without deep statistical expertise.
The decomposable model#
Prophet models a series as a sum of interpretable components,
where \(g(t)\) is the trend (linear, or logistic with saturation \(g(t) = \frac{C}{1 + \exp(-k(t - m))}\), plus automatic changepoints), \(s(t)\) is seasonality (a Fourier series for weekly, yearly or custom cycles), \(h(t)\) captures holidays and events from a supplied list, and \(\varepsilon_t\) is noise.
Strengths and limits#
Prophet is user-friendly (just a ds/y DataFrame), interpretable (each component is
separable), robust to missing data and outliers, detects changepoints automatically, and
scales across many series. Its limits follow from its additive design: it does not model
autoregressive correlations, it suits daily/weekly/monthly business data rather than
high-frequency signals, and it is less powerful than LSTMs or Transformers on complex patterns.
In practice#
from prophet import Prophet
# df has two columns: ds (datestamp) and y (value)
model = Prophet()
model.fit(df)
future = model.make_future_dataframe(periods=90)
forecast = model.predict(future)
model.plot(forecast)
model.plot_components(forecast)
Theme: Signal Processing & Time Series · All terminology
Hint
Mind map — connected ideas
LSTM — Long Short-Term Memory Networks · ARIMA (AutoRegressive Integrated Moving Average) · Time Series · Bayesian Time Series · Signal Processing · Stockout Rate
Hint
More in Signal Processing & Time Series
ARIMA (AutoRegressive Integrated Moving Average) · Bayesian Time Series · Forecast Error · Forecasting Benchmarks · Forecasting Competitions · Log-Space · Low-pass Filtering · LSTM — Long Short-Term Memory Networks · M-Competitions (Makridakis Competitions) · Naïve Baseline Forecast · Seasonal Lag · Seasonality · Signal Processing · Simple Baseline Methods
See also
Source article Adapted (context, re-expressed) in our own words from: Prophet — Time Series Forecasting by Facebook (Meta) (insightful-data-lab.com).