Interactive explorer (Shiny)

The Shiny explorer is a hosted dashboard. It shows the full benchmark results for every method and dataset, without writing code.

Open the Shiny explorer ↗

What you can explore

• scIB metrics

  Batch-correction and bio-conservation scores per method, as bubble tables and overall rankings.

• Integration

  Embeddings, side by side, for each integration scenario.

• Feature selection

  Overlap and rank correlation of the features each method selects.

• Time and memory

  Runtime and peak memory per method.

In Python

mtb.load_results reads the stored tables that ship with the package. They cover the demo datasets only. mtb.plot.bubble draws them, as Load and plot results shows.