Dataswim

A simple api to explore, clean, transform and visualize data. This api is:

• Minimalistic: short names, simple functionalites, minimal parameters
• Pragmatic: focuses on raw efficiency rather than idiomatic code
• Simple stupid: keep it easy to understand for both code and api

Features

• Explore data: describe, search and visualize raw data
• Clean and transform data: select, filter, normalize and reshape data
• Visualize data: many kind of charts and maps for geo data

Install

Using conda:

conda install pandas sqlalchemy seaborn arrow nltk scikit-learn
conda install -c ioam holoviews bokeh
conda install altair --channel conda-forge
pip install --no-cache-dir dataset
pip install pytablewriter goerr gencharts chartjspy stuf \
   deepdish folium influxdb

pip install dataswim --no-deps

Using pip:

pip install --no-cache-dir dataset
pip install dataswim

To get the Altair charts in notebooks running install Vega:

conda install -c conda-forge vega
# or
pip install vega

Documentation

The documentation is available online

Some Jupyter demo notebooks are available as example.

To run the notebooks online click the badge:

Dependencies

To compute data:

• Pandas
• Scikit-learn
• Nltk

To chart data:

• Bokeh with Holoviews,
• Vega Lite with Altair
• Seaborn
• Chartjs
• Folium for the maps

To handle databases:

• Dataset and Sql Alchemy

Supported databases

• Postgresql
• Sqlite
• All those that Sql Alchemy supports

Reporting

To generate static html reports use dsreports

To distribute reports with a http server use django-chartflo