Scinojure Documentation

1 Preface

Scinojure (“Noj”) is an entry point to the Clojure stack for data & science.

It combines a few of the relevant Clojure libraries for data & science and documents common ways of using them together. The included libraries either use tech.ml.dataset directly as tabular data structure or provide high interoperability with it.

Source:

Deps:

Tests:

Note we are using git coordinates at the moment, in order to expose a few relevant features of the current underlying libraries, which are unreleased yet.

Status: Most of the underlying libraries are stable. The experimental parts are marked as such. For some of the libraries, we use a branch for an upcoming release. The main current goal is to provide a clear picture of the direction the stack is going towards, expecting most of it to stabilize soon.

Near term plan - till the end of October 2024

• Work on stabilizing the upcoming releases of the underlying libraries.

• Keep documenting core ideas of the underlying librares and ways to combine them in typical workflows.

• Keep making the docs generate automatic tests using kindly/check.

1.1 Chapters of this book:

• Overview

  • Underlying libraries

  • Other recommended libraries

  • Known issues ❗

• Datasets

  • Datasets

• Machine Learning

  • Machine learning

  • Machine learning specific functionality in tech.ml.dataset

  • AutoML using metamorph pipelines

  • Ordinary least squares with interactions

• Data Visualization

  • Data Visualization with Echarts - DRAFT 🛠

  • Visualizing correlation matrices (experimental) - DRAFT 🛠

• Linear Algebra

  • Intro to Linear Algebra - DRAFT 🛠

source: notebooks/index.clj