scicloj/tablecloth

Dataset manipulation library built on the top of tech.ml.dataset

What it solves

Tablecloth provides a simplified, high-level API for manipulating columnar datasets in Clojure. It acts as a wrapper around the tech.ml.dataset library to make data manipulation more intuitive and consistent, drawing inspiration from popular R data manipulation tools like dplyr, tidyr, and data.table.

How it works

It reorganizes the existing functions of tech.ml.dataset into a streamlined API. Key architectural choices include:

  • Single Entry Point: Common operations are dispatched through a single function based on arguments.
  • Thread-First Support: Functions are designed to enable the use of Clojure's thread-first macro (->) for cleaner data pipelines.
  • Grouped Datasets: It introduces a special dataset type for group-by results, where subsets created after grouping are stored as a column, and most operations are aware of this structure to process data accordingly.

Who it’s for

Data scientists and developers using Clojure who need a powerful and ergonomic way to perform data frame manipulation and cleaning, especially those coming from R or familiar with the Tidyverse ecosystem.

Highlights

  • High-level API for tech.ml.dataset.
  • Support for grouped datasets and specialized aggregation.
  • Designed for thread-first pipeline composition.
  • Inspired by R's dplyr, tidyr, and data.table.

Related

  • Project
    techascent/tech.ml.dataset

    A Clojure library for high-performance tabular data processing on the JVM, providing a functional alternative to Pandas or R's data.table.

  • Project
    hosseinmoein/DataFrame

    A high-performance C++ library for in-memory tabular data analysis, providing efficient data manipulation and statistical tools similar to Pandas but without Python overhead.

  • Project
    huggingface/datasets

    A lightweight library for easy one-line loading and efficient pre-processing of diverse AI-ready datasets (text, audio, image, video) from the Hugging Face Hub or local files.

  • Project
    OpenDCAI/DataFlow

    A data-centric AI system for generating, cleaning, and preparing high-quality LLM training datasets from noisy sources using visual, low-code pipelines.

  • Project
    fluid-cloudnative/fluid

    A Kubernetes-native distributed dataset orchestrator and accelerator that optimizes data access for AI and big data applications by providing dataset abstraction and elastic caching.