Need help with timeseers?
Click the “chat” button below for chat support from the developer who created it, or find similar developers for support.

About the developer

155 Stars 20 Forks MIT License 52 Commits 15 Opened issues


Time should be taken seer-iously

Services available


Need anything else?

Contributors list

# 211,870
31 commits
# 322,734
1 commit
# 555,806
1 commit

Unit Tests


seers - (Noun) plural form of seer - A person who foretells future events by or as if by supernatural means

TimeSeers is an hierarchical Bayesian Time Series model based on Facebooks Prophet, written in PyMC3.

The goal of the TimeSeers project is to provide an easily extensible alternative to Prophet for timeseries modelling when multiple time series are expected to share parts of their parameters.


TimeSeers is designed as a language for building time series models. It offers a toolbox of various components which can be arranged in a formula. We can compose these components in various ways to best fit our problem.

TimeSeers strongly encourages using uncertainty estimates, and will by default use MCMC to get full posterior estimates.

from timeseers import LinearTrend, FourierSeasonality
import pandas as pd

model = LinearTrend() + FourierSeasonality(period=pd.Timedelta(days=365)) + FourierSeasonality(period=pd.Timedelta(days=365))[['t']], data['value'])

Multiplicative seasonality

from timeseers import LinearTrend, FourierSeasonality
import pandas as pd

passengers = pd.read_csv('AirPassengers.csv').reset_index().assign( t=lambda d: pd.to_datetime(d['Month']), value=lambda d: d['#Passengers'] )

model = LinearTrend(n_changepoints=10) * FourierSeasonality(n=5, period=pd.Timedelta(days=365))[['t']], passengers['value'], tune=2000)

model.plot_components(X_true=passengers, y_true=passengers['value']);

Auto-assigning NUTS sampler...
Initializing NUTS using jitter+adapt_diag...
Multiprocess sampling (4 chains in 4 jobs)
NUTS: [sigma, beta, m, delta, k]
Sampling 4 chains, 0 divergences: 100%|██████████| 10000/10000 [00:57<00:00, 173.30draws/s]



PR's and suggestions are always welcome. Please open an issue on the issue list before submitting though in order to avoid doing unnecessary work. I try to adhere to the

style as much as possible. This means:
  • Fitted parameters have a trailing underscore
  • No parameter modification is done in
    methods of model components

We use cookies. If you continue to browse the site, you agree to the use of cookies. For more information on our use of cookies please see our Privacy Policy.