boadata

(B)rowser (O)f (A)rbitrary Data - a Python browser of data. The goal is to create a simple tool for scientists / data miners to browse and plot various data sources in one simple environment.

It's a library, as well as a set of tools.

Most focus is placed on table-like pandas-based objects.

Installation

pipx install boadata

Concepts

• uri - boadata tries to support odo URIs
• data object - wrapper around basic data types (a.k.a odo.resource but with wrapper)
• data tree - browseable that can contain nodes (some of the nodes can be data objects)
• data conversion - between different object types
• view - visual representation of a data objects

CLI tools

Run <command> --help to see full list of options

• boa-describe <uri> - show basic info about a data object
• boa-tree <uri> - list nodes in a data tree
• boa-cat <uri> - print tabular representation of a dataobject
• boa-convert <from+> <to> - convert one data source into another
• boa-table <uri> - a tabular view of data based on textual

Status

• in (slow) development

Requirements

• pandas

• numpy

• numexpr

• click

• tabulate

• blinker (perhaps to be removed)

• h5py (optional)

• sqlalchemy (optional)

• matplotlib (to become optional)

• bokeh (to become optional instead of matplotlib)

• seaborn

• pandas_profiling (optional - dataframe summaries)

• datadotworld (optional - for their datasets)

• Python 3.8+

Supported formats and sources

• File system tree
• HDF5
• CSV (including web links)
• SQL based on SqlAlchemy (sqlite supported)
• pydataset datasets
• seaborn datasets
• MATLAB .fig files
• data.world datasets/tables (that can be imported pandas dataframes)