outliers-filtering 0.1.0

outliers_filtering package for Python-Guide.org

• pip install outliers_filtering

• from outliers_filtering.utils import remove_outliers_numeric, plot_distribution_numeric

Two functions:

remove_outliers_numeric plot_distribution_numeric

Let’s see how to use the functions.

First the function remove_coutliers_numeric, it applies an IQR outliers filtering to normal or lognormal numerical pandas columns.

remove_coutliers_numeric(df,feature,option = “lognormal”,delta=1.5)

• df: is the pandas dataframe where to apply the outlier removal

• feature: name of the column to apply the function

• option: shape of the distribution, either “normal” or “lognormal”

• delta: coefficient to remove outliers outside IQR. default value is 1.5, the higher it is the less outliers you remove.

Example:

Second the function plot_distribution_numeric, it can plot distribution with and without outliers.

plot_distribution_numeric(df, feature, option = ‘lognormal’,title = ‘distribution’, mode = ‘with_outliers’)

• df: is the pandas dataframe where to apply the function

• feature: name of the column to plot the distribution

• option: shape of the distribution, either “normal” or “lognormal”

• title: Title of the plot

• mode: Two possible modes “with_outliers” to plot the raw distribution and “without_outliers” to plot distribution without outliers

Example:

• plot_distribution_numeric(data,’useful_area’, option = ‘lognormal’,title = ‘distribution useful_area’, mode=’with_outliers’)

• plot_distribution_numeric(data, ‘useful_area’, option = ‘lognormal’, title = ‘distribution useful_area without extreme values’, mode=’without_outliers’)