ml_deployment/_3.py at main · ranialaachir/ml_deployment · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
import pandas as pd
import numpy as np
import matplotlib.pyplot as plt

df = pd.read_csv('ENCDM_Menage_2014.csv')

# Sous-groupes niveau 2 de G1 (alimentation)
# et les catégories "signal fort" identifiées
focus = {
    # G1 - Alimentation détaillée
    'Céréales':        'DAP_G01',
    'Viandes':         'DAP_G02',
    'Poissons':        'DAP_G03',
    'Huiles_graisses': 'DAP_G04',
    'Lait_oeufs':      'DAP_G05',
    'Fruits':          'DAP_G06',
    'Légumes':         'DAP_G07',
    'Sucre_confis':    'DAP_G08',
    'Café_thé':        'DAP_G09',
    # G6 - Santé
    'Medicaments':     'DAP_G61',
    'Soins_ambul':     'DAP_G62',
    'Hospitalisation': 'DAP_G63',
    # G9 - Resto/Hotels
    'Restaurants':     'DAP_G91',
    'Hotels_voyages':  'DAP_G92',
    # G7 - Transport
    'Transport_perso': 'DAP_G71',
    'Transport_public':'DAP_G72',
}

print("=== Coefficients budgétaires sous-catégories (%) ===\n")
for label, col in focus.items():
    if col in df.columns:
        df[f'CB_{col}'] = df[col] / df['DAP'] * 100
        means = df.groupby('Quintiles')[f'CB_{col}'].mean()
        variation = means[5.0] - means[1.0]
        direction = "↗" if variation > 0 else "↘"
        print(f"{label:<20} Q1={means[1.0]:.2f}%  Q5={means[5.0]:.2f}%  "
              f"Δ={variation:+.2f}% {direction}")
    else:
        print(f"{label:<20} colonne {col} introuvable")

# Visualisation : Q1 vs Q5 side by side
labels_found = []
q1_vals, q5_vals = [], []

for label, col in focus.items():
    cb_col = f'CB_{col}'
    if cb_col in df.columns:
        means = df.groupby('Quintiles')[cb_col].mean()
        labels_found.append(label)
        q1_vals.append(means[1.0])
        q5_vals.append(means[5.0])

x = np.arange(len(labels_found))
width = 0.35

fig, ax = plt.subplots(figsize=(14, 6))
bars1 = ax.bar(x - width/2, q1_vals, width, label='Q1 (précaires)',
               color='#d62728', alpha=0.85)
bars2 = ax.bar(x + width/2, q5_vals, width, label='Q5 (aisés)',
               color='#9467bd', alpha=0.85)

ax.set_xlabel('Catégorie de dépense')
ax.set_ylabel('% du budget total')
ax.set_title('Structure du panier : Q1 (précaires) vs Q5 (aisés)\nSous-catégories clés')
ax.set_xticks(x)
ax.set_xticklabels(labels_found, rotation=45, ha='right')
ax.legend()
ax.yaxis.set_major_formatter(plt.FuncFormatter(lambda y, _: f'{y:.1f}%'))
plt.tight_layout()
plt.savefig('q1_vs_q5_detail.png', dpi=150, bbox_inches='tight')
plt.show()
print("\nGraphique sauvegardé !")


# === Coefficients budgétaires sous-catégories (%) ===

# Céréales             Q1=9.54%  Q5=4.87%  Δ=-4.66% ↘
# Viandes              Q1=3.06%  Q5=3.16%  Δ=+0.09% ↗
# Poissons             Q1=6.38%  Q5=3.32%  Δ=-3.05% ↘
# Huiles_graisses      Q1=10.85%  Q5=8.00%  Δ=-2.85% ↘
# Lait_oeufs           Q1=1.36%  Q5=1.46%  Δ=+0.10% ↗
# Fruits               Q1=4.95%  Q5=2.32%  Δ=-2.63% ↘
# Légumes              Q1=2.29%  Q5=1.18%  Δ=-1.12% ↘
# Sucre_confis         Q1=2.20%  Q5=2.77%  Δ=+0.57% ↗
# Café_thé             Q1=2.32%  Q5=1.06%  Δ=-1.26% ↘
# Medicaments          Q1=0.91%  Q5=5.69%  Δ=+4.78% ↗
# Soins_ambul          Q1=1.96%  Q5=1.72%  Δ=-0.23% ↘
# Hospitalisation      Q1=1.33%  Q5=2.45%  Δ=+1.12% ↗
# Restaurants          Q1=0.23%  Q5=1.35%  Δ=+1.12% ↗
# Hotels_voyages       Q1=0.71%  Q5=2.33%  Δ=+1.62% ↗
# Transport_perso      Q1=0.20%  Q5=0.59%  Δ=+0.39% ↗
# Transport_public     Q1=0.06%  Q5=0.12%  Δ=+0.06% ↗

# Graphique sauvegardé !