The sinapsis-time-series mono repo provides a comprehensive suite of tools for working with time series data and forecasting models. This release introduce the sinapsis-darts-forecasting package with three specialized templates—Dataframe Loader, Darts Transforms, and Darts Models— that are added as dynamic templates that represent ~60 templates. The release also includes a webapp to simplify the process of loading, transforming, and forecasting time series data. These templates and tools empower developers to efficiently integrate time series forecasting into their applications.
Key Features
-
Dataframe Loader
Converts a pandas DataFrame into a Darts TimeSeries object.
Simplifies the process of preparing time series data for analysis and forecasting.
Supports seamless integration with Darts-based workflows. -
Darts Transforms
Applies several data transformations using Darts transformers to different sources in the time series packet.
Enables developers to preprocess and enhance time series data for improved model performance.
Supports a wide range of transformation techniques for diverse use cases. -
Darts Models
Fits and predicts data inside the container using Darts baseline, statistical, machine learning, and deep learning models.
Provides access to a variety of forecasting models for different time series scenarios.
Simplifies the process of experimenting with and deploying time series forecasting models. -
Web App
A user-friendly web application to make it easier for users to get started with time series forecasting.
Provides an interactive interface for loading data, applying transformations, and training models.
Streamlines the onboarding process for developers and non-technical users alike.