I built this package with the objective is to measure in a simple, fast, but effective way, how the key features of an app/webpage impact over the conversion.
It a feature importance analysis, using BCG (Growth-Share) matrix.
This analysis helps to identify key features, grouping or clustering them into 4 different categories, as the BCG matrix states (see link at the bottom for more information).
To install the package, run this command in the terminal: pip install bcganalysAis
We are going to see the following example.
Important: in this package version, you have to input a dataframe, having as columns: User - Converted - [Features]
So the first column has the users, the second one is the 1 - 0 binary column stating is the user converted or not, and then all the features.
Input dataframe example:
import pandas as pd from bcg_analysis import Generate_BCG # you can find the toy_dataset.csv file in the example folder in the repo df_example = pd.read_csv('toy_dataset.csv',sep=';') df = Generate_BCG(df_example) # generate the plot df.plot_bcg()
# generate the table behind the plot df.df_bcg()
A deeper explanation
About BCG / Growth-Share matrix
Questions / concact
Please send an email to: firstname.lastname@example.org