vizard

Intuitive, Interactive, Easy and Quick Visualizations for Data Science Projects

Installation

pip install vizard

or

pip install git+https://github.com/Ritvik19/vizard.git

Documentation

Instantiate Vizard Object

The Vizard or VizardIn object holds the DataFrame along with its configurations including the PROBLEM_TYPE, DEPENDENT_VARIABLE, CATEGORICAL_INDEPENDENT_VARIABLES, CONTINUOUS_INDEPENDENT_VARIABLES, and TEXT_VARIABLES

import vizard

class config:
    PROBLEM_TYPE = 'regression' or 'classification' or 'unsupervised'
    DEPENDENT_VARIABLE = 'target_variable'
    CATEGORICAL_INDEPENDENT_VARIABLES = [categorical_features]
    CONTINUOUS_INDEPENDENT_VARIABLES = [continuous features]
    TEXT_VARIABLES = [text features]

viz = vizard.Vizard(df, config)
# for interactive plots use:
viz = vizard.VizardIn(df, config)

Exploratory Data Analysis

After Instatiating the Vizard object, you can try different plots for EDA

• Check Missing Values:

  viz.check_missing()
  

• Count of Missing Values:

  viz.count_missing()
  

• Count of Unique Values:

  viz.count_unique()
  

• Count of Missing Values by Group:

  viz.count_missing_by_group(class_variable)
  

• Count of Unique Values by Group: viz.count_unique_by_group(class_variable)

Target Column Analysis

Based on the type of problem, perform a univariate analysis of target column

viz.dependent_variable()

Segmented Univariate Analysis

Based on the type of problem, preform segmented univariate analysis of all feature columns with respect to the target column

• Categorical Variables

    viz.categorical_variables()
  

• Continuous Variables

    viz.continuous_variables()
  

• Text Variables

    viz.wordcloud()
  
    viz.wordcloud_by_group()
  
    viz.wordcloud_freq()
  

Bivariate Analysis

Based on the type of variables, perform bivariate analysis on all the feature columns

• Pairwise Scatter

    viz.pairwise_scatter()
  

• Pairwise Violin

    viz.pairwise_violin()
  

• Pairwise Cross Tabs

    viz.pairwise_crosstabs()
  

Trivariate Analysis

Based on the type of variables, perform trivariate analysis on any of the feature columns

• Trivariate Bubble (Continuous vs Continuous vs Continuous)

    viz.trivariate_bubble(x, y, s)
  

• Trivariate Scatter (Continuous vs Continuous vs Categorical)

    viz.trivariate_scatter(x, y, c)
  

• Trivariate Violin (Categorical vs Continuous vs Categorical)

    viz.trivariate_violin(x, y, c)
  

Correlation Analysis

Based on the type of variables, perform correaltion analysis on all the feature columns

• Correlation Plot

    viz.corr_plot()
  

• Pair Plot

    viz.pair_plot()
  

• Chi Square Plot

    viz.chi_sq_plot()
  

Save the plots to PDF using Viz2PDF

You can also save the plots to a pdf file in order to generate an EDA report

The Viz2PDF object takes in all your Vizard plots and creates a pdf report out of them

viz = vizard.Vizard(df, config)
viz2pdf = vizard.Viz2PDF('viz_report.pdf')

plots = [
    viz.check_missing(),
    viz.count_missing(),
    viz.count_unique(),
    viz.dependent_variable(),
    viz.categorical_variables(),
    viz.continuous_variables(),
    viz.pairwise_scatter(),
    viz.pairwise_violin(),
    viz.pairwise_crosstabs(),
]
viz2pdf(plots)

Usage

1. Classification Case
2. Regression Case
3. Text Classification Case
4. Unsupervised Case
5. Classification Case (Interactive)
6. Regression Case (Interactive)
7. Unsupervised Case (Interactive)