Visualisation

Import des outils / jeu de données

import matplotlib.pyplot as plt
import pandas as pd
import seaborn as sns

from src.config import data_folder
from src.constants import var_categoriques, var_numeriques
from src.utils import init_notebook

init_notebook()

df = pd.read_csv(
    f"{data_folder}/data-cleaned-feature-engineering.csv",
    sep=",",
    index_col="ID",
    parse_dates=True,
)

df.head()

	Year_Birth	Education	Marital_Status	Income	Kidhome	Teenhome	Dt_Customer	Recency	MntWines	MntFruits	...	NumWebVisitsMonth	AcceptedCmp3	AcceptedCmp4	AcceptedCmp5	AcceptedCmp1	AcceptedCmp2	Response	NbAcceptedCampaigns	HasAcceptedCampaigns	NbChildren
ID
5524	1957	Graduation	Single	58138.0	0	0	2012-09-04	58	635	88	...	7	0	0	0	0	0	1	0	False	0
2174	1954	Graduation	Single	46344.0	1	1	2014-03-08	38	11	1	...	5	0	0	0	0	0	0	0	False	2
4141	1965	Graduation	Together	71613.0	0	0	2013-08-21	26	426	49	...	4	0	0	0	0	0	0	0	False	0
6182	1984	Graduation	Together	26646.0	1	0	2014-02-10	26	11	4	...	6	0	0	0	0	0	0	0	False	1
5324	1981	PhD	Married	58293.0	1	0	2014-01-19	94	173	43	...	5	0	0	0	0	0	0	0	False	1

5 rows × 28 columns

Variables globales

var_categoriques_extra = ["NbAcceptedCampaigns", "HasAcceptedCampaigns", "NbChildren"]

var_categoriques_fe = (
    var_categoriques + var_categoriques_extra
)  ## todo: sauvegarder toutes ces listes dans un pd.Series et les lire

df[var_categoriques_fe] = df[var_categoriques_fe].astype(str).astype("category")

Visualisation

Relation avec la variable cible (Response)

for var in var_numeriques:
    _, ax = plt.subplots(1, 2, figsize=(10, 3))
    sns.boxplot(df, x=df[var], y=df["Response"], width=0.25, ax=ax[0])
    sns.histplot(
        df,
        x=df[var],
        kde=True,
        ax=ax[1],
        hue=df["Response"],
        stat="probability",
        common_norm=False,
    )
    plt.show()

for var in var_categoriques_fe:
    _, ax = plt.subplots(1, 2, figsize=(10, 3))

    sns.histplot(
        df,
        x=df[var],
        hue=df["Response"],
        multiple="dodge",
        shrink=0.5,
        ax=ax[0],
    )
    sns.histplot(
        df,
        hue=df[var],
        x=df["Response"],
        multiple="dodge",
        shrink=0.5,
        ax=ax[1],
    )

    plt.show()

Graphiques supplémentaires

_, ax = plt.subplots(1, 2, figsize=(12, 4))

ax[0].set_title("Revenu en fonction du nombre de campagnes acceptées")

sns.boxplot(y=df["Income"], x=df["NbAcceptedCampaigns"], ax=ax[0])
sns.histplot(
    df,
    x="Income",
    hue="NbAcceptedCampaigns",
    kde=True,
    stat="probability",
    common_norm=False,
    ax=ax[1],
)

<Axes: xlabel='Income', ylabel='Probability'>

_, ax = plt.subplots(1, 2, figsize=(12, 4))

ax[0].set_title("Revenu en fonction du nombre de campagnes acceptées")

sns.boxplot(y=df["Income"], x=df["HasAcceptedCampaigns"], ax=ax[0])
sns.histplot(
    df,
    x="Income",
    hue="HasAcceptedCampaigns",
    kde=True,
    stat="probability",
    common_norm=False,
    ax=ax[1],
)

<Axes: xlabel='Income', ylabel='Probability'>

todo

sns.histplot(
    data=df, x="Education", hue="HasAcceptedCampaigns", multiple="dodge", shrink=0.8
)

<Axes: xlabel='Education', ylabel='Count'>

sns.histplot(
    data=df,
    x="NbChildren",
    hue="Response",
    multiple="dodge",
    discrete=True,
    shrink=0.4,
)

<Axes: xlabel='NbChildren', ylabel='Count'>

sns.histplot(
    data=df,
    x="NbChildren",
    hue="HasAcceptedCampaigns",
    multiple="dodge",
    discrete=True,
    shrink=0.4,
)

<Axes: xlabel='NbChildren', ylabel='Count'>

Matrice de corrélation

plt.figure(figsize=(12, 12))
sns.heatmap(
    df.corr()[df.corr().abs() > 0.5],
    annot=True,
    cmap="BrBG",
    linewidths=0.5,
    vmax=1,
    vmin=-1,
)

/tmp/ipykernel_2065/1235220098.py:3: FutureWarning: The default value of numeric_only in DataFrame.corr is deprecated. In a future version, it will default to False. Select only valid columns or specify the value of numeric_only to silence this warning.
  df.corr()[df.corr().abs() > 0.5],

<Axes: >

todo

## Exemple de graphique sur plusieurs colonnes
## sns.displot(
#     data=df,
#     x="Marital_Status",
#     hue="Response",
#     col="Kidhome",
#     multiple="dodge",
#     shrink=0.8,
## )