Optimizing Alluvial Plots

Sort k-partite graphs with node order, layer order, and node grouping optimized with a heuristic to (nearly) minimize edge crossings. Useful for improving visualizations with alluvial plots by "untangling" the graphs.

wompwomp

wompwomp solves the Weighted (permutation) Optimization of Multiple Partitions-Weighted (label) Optimization of Multiple Partitions (W_POMP--W_LOMP) problem.

alt text

wompwomp functions/commands

alt text

Installation:

CRAN - Requires system R to be installed

install.packages("wompwomp")

GitHub

if (!require("remotes", quietly = TRUE))
    install.packages("remotes")
remotes::install_github("pachterlab/wompwomp")

Getting Started

See the vignettes for wompwomp in the "vignettes" folder.

wompwomp_intro.Rmd introduces the package and its functions, and provides a walkthrough of how to use wompwomp on simulated data.
plotting.Rmd visualizes the results of wompwomp with ggplot2 and ggalluvial.

Usage

The I/O for each of wompwomp's functions is as follows:

prep_for_lodes: dataframe, csv, or tibble (grouped or ungrouped) --> dataframe (grouped)
sort_to_uncross: dataframe, csv, or tibble (grouped or ungrouped) --> dataframe (grouped)
compute_crossing_objective: dataframe, csv, or tibble (grouped or ungrouped) --> list

The input table can have one of two formats:

Ungrouped: columns specified by graphing_columns, where each row corresponds to a separate entity
Grouped: columns specified by graphing_columns and column_weights, where each row corresponds to a combination of graphing_columns, and column_weights specified the number of items in this combination

Read our preprint on arXiv here. Note that some functionalities and names have been changed since the preprint. To download the version of wompwomp reflected in the preprint, please see the "arxiv_v1" branch of this repository. Reproduce figures from our preprint at https://github.com/pachterlab/ROP_2025.