jupyter-scatter 0.1.0

A scatter plot extension for Jupyter Notebook and Lab

jscatter

A lightweight but scalable scatter plot extension for Jupyter Lab and Notebook

Why? After embedding data we want to explore the embedding space, which typically involves three things besides plotting the data as a 2D scatter. First, we want to interactively adjust the view (e.g., via panning & zooming) and the visual point encoding (e.g., the point color, opacity, or size). Second, we want to be able to select/highlight points. And third, we want to compare multiple embeddings (e.g., via animation, color, or point connections). The goal of jscatter is to support all three requirements and scale to millions of points.

How? Internally, jscatter is using regl-scatterplot

---

Development

Requirements:

• Conda >= 4.8

Installation:

git clone https://github.com/flekschas/jscatter/ jscatter && cd jscatter
conda env create -f environment.yml && conda activate jscatter
pip install -e .

Enable the Notebook Extension:

jupyter nbextension install --py --symlink --sys-prefix jscatter
jupyter nbextension enable --py --sys-prefix jscatter

Enable the Lab Extension:

jupyter labextension develop jscatter

After Changing Python code: simply restart the kernel.

After Changing JavaScript code: do cd js && npm run build and reload the browser tab.