TaGra: an open Python package for easily generating graphs from data tables through manifold learning

Area of Science:

• Data analysis
• Bioinformatics
• Computational chemistry

Background:

• High-dimensional data analysis is crucial across scientific fields.
• Traditional dimensionality reduction methods can distort data, hindering interpretation and visualization.
• Visualizing relationships and similarities in tabular data remains a challenge.

Purpose of the Study:

• Introduce TaGra, an accessible software package for generating similarity graphs from tabular data.
• Provide a novel approach to visualize and analyze high-dimensional datasets.
• Facilitate the identification of patterns, outliers, and class separations within data.

Main Methods:

• TaGra constructs a graph representing similarity relations from input tabular data.
• The package offers functionalities for 2D data visualization.
• It includes methods for identifying typical data points and outliers.

Main Results:

• TaGra successfully visualizes complex datasets in an interpretable 2D space.
• The software aids in distinguishing between different data classes and identifying anomalies.
• Demonstrates effective visualization of mutual distances and similarity patterns.

Conclusions:

• TaGra offers an effective solution for visualizing and analyzing high-dimensional tabular data.
• The package enhances the interpretability of complex datasets across various scientific disciplines.
• TaGra is readily available for use and further development.