Diviner Documentation

Diviner is an open source library for large-scale (multi-series) time series forecasting. It serves as a wrapper around popular open source forecasting libraries, providing a consolidated framework for simplified modeling of many discrete series components with an efficient high-level API.

To get started with Diviner, see the quickstart guide (Quickstart Guide).

Individual forecasting library API guides can be found here:

• Grouped Prophet API

• Grouped Pmdarima API

Examples of each of the wrapped forecasting libraries can be found here.

Contents:

• Diviner: Grouped Timeseries Forecasting at scale
  • Is this right for my project?
  • Grouped Modeling Wrappers
  • Installing
  • Documentation
  • Community & Contributing
• Quickstart Guide
  • Installing Diviner
  • Downloading Examples
  • Input Dataset Structure
  • Basic Example
• Tutorials and Examples
  • GroupedProphet Notebook
  • GroupedProphet Scripts
  • GroupedPmdarima Notebook
  • GroupedPmdarima Scripts
• Grouped Prophet
  • Grouped Prophet API
  • Utilities
  • Class Signature
• Grouped pmdarima
  • Grouped pmdarima API
  • Utilities
  • Class Signature of GroupedPmdarima
  • Grouped pmdarima Analysis tools
  • Class Signature of PmdarimaAnalyzer
• Data Processing
  • Pandas DataFrame Group Processing API
  • Developer API for Data Processing
• Contributing Guide
  • Communication
  • Coding Style
  • Code formatting
  • Sign your work
• Changelog
  • 0.1.1 - Support pmdarima 2.x
  • 0.1.0 - Initial release