Visualizing balances of compositional data: A new alternative to balance dendrograms.

Area of Science:

• Statistics
• Bioinformatics
• Data Science

Background:

• Balances are crucial in compositional data analysis but challenging to conceptualize, especially with high-dimensional data.
• The balance dendrogram is a common visualization tool but lacks intuitiveness and scalability for large datasets.

Purpose of the Study:

• Introduce the R package 'balance' as a novel visualization method for compositional data balances.
• Provide an intuitive and scalable alternative to the balance dendrogram for high-dimensional data analysis.

Main Methods:

• Developed the 'balance' R package for visualizing compositional data.
• Implemented a projection onto a common scale to facilitate direct comparisons.
• Demonstrated visualization of subsets of balances without disrupting interpretation.

Main Results:

• The 'balance' package offers an alternative visualization that retains information from the balance dendrogram.
• This method projects data onto a common scale, enhancing interpretability and comparison.
• The package effectively visualizes balances for high-dimensional data, as shown with a meta-genomics dataset.

Conclusions:

• The 'balance' R package provides an intuitive and scalable solution for visualizing compositional data balances.
• This tool facilitates the analysis of complex, high-dimensional datasets, improving upon existing methods like the balance dendrogram.