Timeframes – Time Series Forecasting Utilities

Getting Started

Install timeframes and start validating your time series models in just a few minutes.

Installation

pip install timeframes

Quick Start

Import timeframes and integrate it with your favorite machine learning library:

import timeframes as ts
from sklearn.linear_model import LinearRegression

# Load example data
df = ts.load_example("air_passengers")

# Split into train, validation, and test sets
train, val, test = ts.split(df, ratios=(0.7, 0.2, 0.1))

# Validate using walk-forward cross-validation
model = LinearRegression()
report = ts.validate(
    model,
    df,
    target_col="AirPassengers",
    method="walkforward",
    mode="expanding",
    folds=5
)

print(report)
# {'mae': 0.213, 'rmse': 0.322, 'smape': 3.9}

Core Concepts

Time Series Splitting

Use ts.split() to divide your dataset into train, validation, and test sets while preserving chronological order.

Cross-Validation

Use ts.validate() for walk-forward or temporal K-Fold validation to assess model consistency across time windows.

Backtesting

Use ts.backtest() to evaluate your model on truly unseen data, simulating real-world forecasting conditions.