A python package to help Data Scientists, Machine Learning Engineers and Analysts better understand data. Gives quick insights about given data; general dataset statistics, shape of dataset, number of unique data types, number of numerical and non-numerical columns, missing data statistics, missing data heatmap and provides methodology to impute missing data.
Project description
datastand
Why datastand? Data + Understand
A python package to help Data Scientists, Machine Learning Engineers and Analysts better understand data. Gives quick insights about a given dataset.
Installation
Run the following command on the terminal to install the package:
pip install datastand
Usage :
Code:
from datastand import datastand
import pandas as pd
df = pd.read_csv("path/to/target/dataframe")
datastand(df)
Output:
General stats:
==================
Shape of DataFrame: (1202, 13)
Number of unique data types : {dtype('int64'), dtype('O')}
Number of numerical columns: 2
Number of non-numerical columns: 11
Missing data:
=======================
DataFrame contains 2670 missing values (17.09%) as follows column-wise:
-----------------------------------------------------------------------
Gender 41
Car_Category 372
Subject_Car_Colour 697
Subject_Car_Make 248
LGA_Name 656
State 656
dtype: int64
-----------------------------------------------------------------------
Do you wish to long-list missing data statistics?(y/n): y
.
.
.
Code:
# This function is already available in the DataStand class and also available separately
# Here we're running it separately
from datastand import plot_missing
plot_missing(df)
Output:
Code:
from datastand import impute_missing
impute_missing(df)
Output:
Imputing missing data...
100%|████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 80/80 [00:02<00:00, 30.52it/s]
Imputation complete.
Author/Maintainer
Vincent N. [LinkedIn] [Twitter]
Project details
Release history Release notifications | RSS feed
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
Built Distribution
Hashes for datastand-2.5.0-py3-none-any.whl
Algorithm | Hash digest | |
---|---|---|
SHA256 | a452302ffe1863d657a86dbefddb9ad0fdd341fd56734009d267fef68b7309cc |
|
MD5 | ff31a7f64b7c192d6972ed633afcc4ed |
|
BLAKE2b-256 | a5813622337a30f0990fa3e4d19ac4b193ed12289b730a338c647ece6b954cd4 |