Sankey Plot App

Interactive Shiny app that transforms a wide Excel timepoint table into an explorable Sankey diagram, with PDF/HTML export.

Structure

sankey_plot/
│
├── sankey_app.Rproj          # RStudio project file
├── .gitignore                # Files ignored by Git
├── global.R                  # Package loading, theme configuration, shared functions
├── ui.R                      # User interface definition
├── server.R                  # Server logic, Sankey generation, download handlers
├── R/
│   └── data_processing.R     # Data processing functions (Excel → Sankey nodes and links)
├── www/
│   └── styles.css            # Custom CSS styles loaded automatically by Shiny
├── data/
│   └── example_1.xlsx
└── README.md

Running the app

Option 1: Run directly from GitHub

The application can be launched automatically from R without manually downloading the repository.

Install Shiny if needed:

install.packages("shiny")

Then run:

shiny::runGitHub(
  repo = "sankey_plot",
  username = "BiMaU-PCCB"
)

This command downloads the latest version of the repository from GitHub and opens the Shiny application automatically.

---

Option 2: Run locally

From R, set your working directory to the sankey_app/ folder:

shiny::runApp()

Or, from anywhere:

shiny::runApp("path/to/sankey_app")

In RStudio, open any of global.R, ui.R, or server.R and click Run App. RStudio will automatically detect the Shiny application structure.

Required Excel structure

The input Excel file must follow this structure:

• Data must be stored in the first worksheet.
• First column: patient ID (one row per patient).
• Remaining columns: one column per timepoint (e.g. Baseline, Month 3, Month 6).
• Each cell must contain a consistent categorical value representing the state/category at that timepoint.
• Column names must not contain _, $, or @.
• Missing values must be represented as empty cells (not NA or NaN).

An example Excel file (Example.xlsx) is included in the data/ folder and can be used to test the application. This file follows the required input structure.

This structure is also described inside the application in the "Info: Data structure" button.

Dependencies

The application requires the following R packages:

• bslib
• colourpicker
• htmlwidgets
• jsonlite
• magick
• networkD3
• RColorBrewer
• readxl
• shiny
• webshot2

Install any missing packages from R with:

install.packages(c(
  "bslib",
  "colourpicker",
  "htmlwidgets",
  "jsonlite",
  "magick",
  "networkD3",
  "RColorBrewer",
  "readxl",
  "shiny",
  "webshot2"
))