Project description

An extension to Pandas' describe function. It operates on a Pandas dataframe to produce a dataframe of stats for numerical columns and another dataframe of stats for categorical columns.

Numerical columns are selected using:

df.select_dtypes(include=['number'])

While categorical columns are selected using:

df.select_dtypes(include=['object'])

See the numpy dtype hierarchy.

Numerical column stats:

• Type
• Density
• No. of Values
• No. of Unique Values
• No. of NaN
• No. of Zeros
• No. of +ve Values
• Min
• Max
• Mean
• Std
• Median
• Mode
• Skew
• Kurtosis
• No. of 3 Sigma Outliers

Object column stats:

• Type
• Density
• No. of Values
• No. of Unique Values No. of NaN

Project Files

• datasets: A folder containing a set of CSV datasets.
• describe2.py: Describe2 Python class file.
• examples.ipynb: A Jupyter notebook demonstrating Describe2 on sample datasets.
• README.md: This file. It contains information about the project.
• LICENSE.txt: GNU GPL v3.0 License file.

Usage

Import Describe2:

import Describe2
describe2 = Describe2.Describe2()

Generate stats for a dataset:

cars = pd.read_csv('cars.csv')
num, obj = describe2.d2(cars)

num is a dataframe holding the numerical columns stats, and obj is a dataframe holding the categorical columns stats.

License

This is an open source free program provided under Version 3 of the GNU GENERAL PUBLIC LICENSE. A copy of the license is available in LICENSE.txt at the root of the source code. If not, please see http://www.gnu.org/licenses/.