compotime 0.1.0

A library for forecasting compositional time series

compotime

compotime is a library for forecasting compositional time series in Python. At the moment, it provides an implementation of the models described in the paper "Forecasting compositional time series: A state space approach" (Snyder, R.D. et al, 2017). It is constantly tested to be compatible with the major machine learning and statistics libraries within the Python ecosystem.

Basic usage

This example uses adapted data on the global share of energy consumption by source (1965-2021).

import pandas as pd

from compotime import LocalTrendForecaster, preprocess

URL = "https://raw.githubusercontent.com/mateuja/compotime/main/examples/data/share_energy_source.csv"

date_parser = lambda x: pd.Period(x, "Y")
time_series = (
  pd.read_csv(URL, parse_dates=["Year"], date_parser=date_parser)
  .set_index("Year")
  .pipe(preprocess.treat_small, 0.001)
)

model = LocalTrendForecaster()
model.fit(time_series)
model.predict(horizon=10)

For more details, see the Documentation.