Datui is a TUI application I created to make viewing large tabular datasets in Parquet, CSV, and more quick and easy (hive partitioning too). It is also a landing page for your configurable catalog of data, creates exportable charts, has built in data-quality checks, and lets you query and transform with SQL. Try it as a python module to make viewing/debugging DataFrames in a terminal session easy. It's built on Polars LazyFrames, so only the rows on screen are read.
Additionally, Datui can read custom binary formats (via a toml spec), CAN, FIX, Excel, Arrow, and others.
You can install it with pip install datui, brew, winget, package managers, or a curl script.
Try opening up public parquet with this:
datui s3://noaa-ghcn-pds/parquet/by_year/YEAR=2024/
Which data formats or workflows would you like to be able to inspect?
derekwisong•1h ago
This started because I wanted a nice visual debugger for Polars dataframes, while working on analysis of some aircraft engine data. What started as a tool for a quick look, eventually became a place to add all the things I usually open a notebook for when starting work on a new dataset: charting, analysis, diving into groups.
This is a single-binary terminal viewer for tabular data: local or remote. Built on Rust, Ratatui, and Polars. Directories (and bucket prefixes) can be opened as one table.
I took a lot of inspiration from VisiData, but wanted to focus on viewing and not editing (datui does support exporting) large datasets.
I also wanted it to be themeable, have mouse support, and a customizable home page that I can use as a data catalog.