https://github.com/dabl/dabl

Data Analysis Baseline Library
https://github.com/dabl/dabl

Last synced: about 2 months ago
JSON representation

Data Analysis Baseline Library

• Host: GitHub
• URL: https://github.com/dabl/dabl
• Owner: dabl
• License: bsd-3-clause
• Created: 2018-09-14T19:11:47.000Z (over 6 years ago)
• Default Branch: main
• Last Pushed: 2024-08-07T21:38:41.000Z (5 months ago)
• Last Synced: 2024-10-07T22:21:42.952Z (2 months ago)
• Language: Jupyter Notebook
• Homepage: https://dabl.github.io/
• Size: 7.4 MB
• Stars: 724
• Watchers: 23
• Forks: 105
• Open Issues: 89
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

• StarryDivineSky - dabl/dabl

README

        
# dabl

[![CI](https://github.com/dabl/dabl/actions/workflows/ci.yml/badge.svg)](https://github.com/dabl/dabl/actions/workflows/ci.yml)

The data analysis baseline library.

- "Mr Sanchez, are you a data scientist?"

- "I dabl, Mr president."

Find more information on the [website](https://dabl.github.io/).

## Try it out

```

pip install dabl

```

or [![Binder](https://mybinder.org/badge_logo.svg)](https://mybinder.org/v2/gh/dabl/dabl/main)

## Current scope and upcoming features

This library is very much still under development. Current code focuses mostly on exploratory visualization and preprocessing.

There are also drop-in replacements for GridSearchCV and RandomizedSearchCV using successive halfing.

There are preliminary portfolios in the style of

[POSH

auto-sklearn](https://ml.informatik.uni-freiburg.de/papers/18-AUTOML-AutoChallenge.pdf)

to find strong models quickly.  In essence that boils down to a quick search

over different gradient boosting models and other tree ensembles and

potentially kernel methods.

Check out the [the website](https://dabl.github.io/dev/) and [example gallery](https://dabl.github.io/0.1.9/auto_examples/index.html) to get an idea of the visualizations that are available.

Stay Tuned!

## Related packages

## Lux

[Lux](https://github.com/lux-org/lux) is an awesome project for easy interactive visualization of pandas dataframes within notebooks.

### Pandas Profiling

[Pandas Profiling](https://github.com/pandas-profiling/pandas-profiling) can

provide a thorough summary of the data in only a single line of code. Using the

```ProfileReport()``` method, you are able to access a HTML report of your data

that can help you find correlations and identify missing data.

`dabl` focuses less on statistical measures of individual columns, and more on

providing a quick overview via visualizations, as well as convienient

preprocessing and model search for machine learning.