A python library for monitoring, comparing and extracting insights from data.


Keywords
metrics, data, evaluation, data-analysis, data-science, framework, insights, monitoring, python
License
BSD-1-Clause
Install
pip install insight==1.0

Documentation

🧿 insight

GitHub top language GitHub PyPI - Downloads GitHub Repo stars

A python package to quickly understand, assess, and compare pandas Series/DataFrames.

The predominant functions in the package focus on easy-to-use metrics and intelligent plotting functions. The metrics can also be configured from YAML to allow for simple to configure benchmarking and assessment scripts.

PyPI CodeQL Status CI Status Coverage Code Smells pre-commit.ci status

Installation

pip install insight

Usage

Metrics

At the core of insight are the metrics classes which can be evaluated on one series, two series, one dataframe or two dataframes.

>>> import insight.metrics as m
>>> metric = m.EarthMoversDistance()
>>> metric(df['A'], df['B'])
0.14

Plotting

The package provides various plotting functions which allow you to easily explore any series, dataframe or multiple dataframes.

>>> import insight.plotting as p
>>> p.plot_dataset([df1, df2])

Migrations

insight populates the results to the Postgres database configured by environment variables. To run migrations against it, simply:

insight-migrations
distribution plots