If you have hundreds of csvs you need to quickly digest and understand, you can use crystal-ball to help with the onboarding and data exploration phase of your project. You will be able to immediately dive into an expansive data set without getting lost.

CrystalBall Features

• find specific columns and tables you may be interested in, but may have otherwise not known where to look.

• walk through connections between different csvs,

• compare and establish foreign key and primary key relationships by using simple boxplots

• dynamically create a master table of useful information while simultaneously recording your step-by-step process for future reference.

Installation and Usage

pip install crystal-ball</programlisting> You can start using CrystalBall right away by importing it and initializing it with a relative directory containing the CSVs.

import crystalball as cb

ball = cb.run("insert relative directory here")

Methods

Note that all methods that involve searching via keynames are case sensitive.

cb.contains(keywords: list, all_colnames: list=None) → list

Check if keywords exist in all_colnames.

cb.featureSearch(keywords: list, all_colnames: list=None) → list

Find the columns that contain the keywords.

cb.tableSearch(keywords: list, csvname_to_colnames_list=None, mode: str=UNION) → list

Find the tables that contain the keywords.

cb.openTable(rel_dir: str, indices: list=[0]) → DataFrame

Open the csv that is referenced by the given relative directory.

cb.subTable(supertable: DataFrame, chosen_index: list, chosen_columns: list) → DataFrame

Create a subtable from a supertable.

cb.mergeTables(tables: list) → DataFrame

Sequentially merge a list of tables that all share a common index.

cb.analyzeRelationships(to_analyze: list, visualize: bool=True) → DataFrame

Analyze basic stats of one or more different Series.

compareRelationship(to_analyze1: Series, to_analyze2: Series, visualize: bool=False) → DataFrame

Compare and contrast the difference between two Series.

cb.export(df_to_export: DataFrame, write_to: str, export_type: str=CSV) → None

Exports contents of dataframe to relative location specified by write_to arg.