Daany -- DAta ANalYtics on .NET

TL;DR

Daany is a cross-platform, .NET-based data analytics library in C# that provides data preparation, feature engineering, and linear algebra functionalities, similar to Python's Pandas, with support for time series and optimization methods.

Contribution

It introduces a comprehensive data analytics library for .NET, including DataFrame, time series decomposition, and linear algebra features, with detailed implementation and usage guidance.

Findings

Supports .NET Standard 2.1 and .NET Core 3.0+

Includes time series decomposition and optimization methods

Provides a Pandas-like DataFrame implementation

Abstract

Daany is .NET and cross platform data analytics and linear algebra library written in C# supposed to be a tool for data preparation, feature engineering and other kind of data transformations and feature engineering. The library is implemented on top of .NET Standard 2.1 and supports .NET Core 3.0 and above separated on several Visual Studio projects that can be installed as a NuGet package. The library implements DataFrame as the core component with extensions of a set of data science and linear algebra features. The library contains several implementation of time series decomposition (SSA, STL ARIMA), optimization methods (SGD) as well as plotting support. The library also implements set of features based on matrix, vectors and similar linear algebra operations. The main part of the library is the Daany.DataFrame with similar implementation that can be found in python based Pandas library. The paper presents the main functionalities and the implementation behind the Daany packages in the form of developer guide and can be used as manual in using the Daany in every-day work. To end this the paper shows the list of papers used the library.