File size: 10,835 Bytes
5e2aaa0
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
"""
Clustering Analysis Module
=========================

This module implements various clustering algorithms for customer segmentation.
"""

import numpy as np
import pandas as pd
from sklearn.cluster import KMeans, DBSCAN
from sklearn.metrics import silhouette_score, calinski_harabasz_score
import streamlit as st

class ClusteringAnalyzer:
    """
    Handles clustering analysis for customer segmentation.
    """
    
    def __init__(self):
        self.kmeans_model = None
        self.dbscan_model = None
        self.optimal_clusters = None
        self.cluster_labels = {}
        
    def find_optimal_clusters(self, scaled_data, max_clusters=10):
        """Find optimal number of clusters using multiple methods."""
        if scaled_data is None:
            st.error("No scaled data available. Please preprocess data first.")
            return None
        
        cluster_range = range(2, max_clusters + 1)
        inertias = []
        silhouette_scores = []
        calinski_scores = []
        
        progress_bar = st.progress(0)
        status_text = st.empty()
        
        for i, k in enumerate(cluster_range):
            status_text.text(f'Evaluating {k} clusters...')
            
            kmeans = KMeans(n_clusters=k, random_state=42, n_init=10)
            cluster_labels = kmeans.fit_predict(scaled_data)
            
            inertias.append(kmeans.inertia_)
            silhouette_scores.append(silhouette_score(scaled_data, cluster_labels))
            calinski_scores.append(calinski_harabasz_score(scaled_data, cluster_labels))
            
            progress_bar.progress((i + 1) / len(cluster_range))
        
        status_text.text('Optimization complete!')
        
        # Find optimal clusters based on silhouette score
        optimal_silhouette = cluster_range[np.argmax(silhouette_scores)]
        optimal_calinski = cluster_range[np.argmax(calinski_scores)]
        
        # Store results
        self.optimization_results = {
            'cluster_range': list(cluster_range),
            'inertias': inertias,
            'silhouette_scores': silhouette_scores,
            'calinski_scores': calinski_scores,
            'optimal_silhouette': optimal_silhouette,
            'optimal_calinski': optimal_calinski
        }
        
        self.optimal_clusters = optimal_silhouette
        
        st.success(f"โœ… Optimal clusters found: {self.optimal_clusters} (based on Silhouette Score)")
        
        return self.optimization_results
    
    def apply_kmeans(self, scaled_data, n_clusters=None):
        """Apply K-Means clustering."""
        if scaled_data is None:
            st.error("No scaled data available. Please preprocess data first.")
            return None
        
        if n_clusters is None:
            n_clusters = self.optimal_clusters or 5
        
        with st.spinner(f'Applying K-Means clustering with {n_clusters} clusters...'):
            self.kmeans_model = KMeans(n_clusters=n_clusters, random_state=42, n_init=10)
            kmeans_labels = self.kmeans_model.fit_predict(scaled_data)
        
        # Calculate metrics
        silhouette_avg = silhouette_score(scaled_data, kmeans_labels)
        calinski_score = calinski_harabasz_score(scaled_data, kmeans_labels)
        
        self.cluster_labels['kmeans'] = kmeans_labels
        
        results = {
            'labels': kmeans_labels,
            'n_clusters': n_clusters,
            'silhouette_score': silhouette_avg,
            'calinski_score': calinski_score,
            'inertia': self.kmeans_model.inertia_,
            'centers': self.kmeans_model.cluster_centers_
        }
        
        st.success(f"โœ… K-Means clustering completed!")
        st.info(f"Silhouette Score: {silhouette_avg:.3f} | Calinski-Harabasz Score: {calinski_score:.3f}")
        
        return results
    
    def apply_dbscan(self, scaled_data, eps=0.5, min_samples=5):
        """Apply DBSCAN clustering."""
        if scaled_data is None:
            st.error("No scaled data available. Please preprocess data first.")
            return None
        
        with st.spinner(f'Applying DBSCAN clustering (eps={eps}, min_samples={min_samples})...'):
            self.dbscan_model = DBSCAN(eps=eps, min_samples=min_samples)
            dbscan_labels = self.dbscan_model.fit_predict(scaled_data)
        
        # Calculate metrics
        n_clusters = len(set(dbscan_labels)) - (1 if -1 in dbscan_labels else 0)
        n_noise = list(dbscan_labels).count(-1)
        
        self.cluster_labels['dbscan'] = dbscan_labels
        
        results = {
            'labels': dbscan_labels,
            'n_clusters': n_clusters,
            'n_noise': n_noise,
            'eps': eps,
            'min_samples': min_samples
        }
        
        # Calculate silhouette score only if we have more than 1 cluster and non-noise points
        if n_clusters > 1:
            non_noise_mask = dbscan_labels != -1
            if np.sum(non_noise_mask) > 1:
                silhouette_avg = silhouette_score(scaled_data[non_noise_mask], 
                                                dbscan_labels[non_noise_mask])
                results['silhouette_score'] = silhouette_avg
        
        st.success(f"โœ… DBSCAN clustering completed!")
        st.info(f"Clusters: {n_clusters} | Noise points: {n_noise}")
        
        return results
    
    def analyze_clusters(self, data, algorithm='kmeans'):
        """Analyze cluster characteristics."""
        # Normalize algorithm name
        algo_key = algorithm.lower().replace('-', '').replace(' ', '')
        
        if algo_key not in self.cluster_labels:
            st.error(f"No {algorithm} clustering results found. Please run clustering first.")
            return None
        
        cluster_labels = self.cluster_labels[algo_key]
        
        # Create consistent column name (use the format that actually gets created)
        if algo_key == 'kmeans':
            cluster_col = 'Kmeans_Cluster'  # Match what we see in the error
        elif algo_key == 'dbscan':
            cluster_col = 'DBSCAN_Cluster'
        else:
            cluster_col = f'{algorithm}_Cluster'
        
        # Add cluster labels to data
        analysis_data = data.copy()
        analysis_data[cluster_col] = cluster_labels
        
        # Calculate cluster statistics
        numeric_cols = analysis_data.select_dtypes(include=[np.number]).columns
        numeric_cols = [col for col in numeric_cols if not col.endswith('_Cluster')]
        
        cluster_stats = analysis_data.groupby(cluster_col)[numeric_cols].agg(['mean', 'std', 'count'])
        
        # Calculate spending analysis if available
        spending_analysis = None
        if 'Spending Score (1-100)' in analysis_data.columns:
            spending_analysis = analysis_data.groupby(cluster_col)['Spending Score (1-100)'].agg(['mean', 'std', 'min', 'max', 'count'])
        
        results = {
            'data_with_clusters': analysis_data,
            'cluster_stats': cluster_stats,
            'spending_analysis': spending_analysis,
            'cluster_distribution': analysis_data[cluster_col].value_counts().sort_index()
        }
        
        return results
    
    def get_cluster_profiles(self, data, algorithm='kmeans'):
        """Generate customer profiles for each cluster."""
        # Normalize algorithm name
        algo_key = algorithm.lower().replace('-', '').replace(' ', '')
        
        if algo_key not in self.cluster_labels:
            return None
        
        cluster_labels = self.cluster_labels[algo_key]
        
        # Create consistent column name (use the format that actually gets created)
        if algo_key == 'kmeans':
            cluster_col = 'Kmeans_Cluster'  # Match what we see in the error
        elif algo_key == 'dbscan':
            cluster_col = 'DBSCAN_Cluster'
        else:
            cluster_col = f'{algorithm}_Cluster'
        
        analysis_data = data.copy()
        analysis_data[cluster_col] = cluster_labels
        
        profiles = []
        
        for cluster in sorted(analysis_data[cluster_col].unique()):
            if cluster == -1:  # Skip noise points in DBSCAN
                continue
                
            cluster_data = analysis_data[analysis_data[cluster_col] == cluster]
            
            profile = {
                'cluster': cluster,
                'size': len(cluster_data),
                'percentage': len(cluster_data) / len(analysis_data) * 100
            }
            
            # Add feature statistics
            if 'Age' in cluster_data.columns:
                profile['avg_age'] = cluster_data['Age'].mean()
                profile['age_std'] = cluster_data['Age'].std()
            
            if 'Annual Income (k$)' in cluster_data.columns:
                profile['avg_income'] = cluster_data['Annual Income (k$)'].mean()
                profile['income_std'] = cluster_data['Annual Income (k$)'].std()
            
            if 'Spending Score (1-100)' in cluster_data.columns:
                profile['avg_spending'] = cluster_data['Spending Score (1-100)'].mean()
                profile['spending_std'] = cluster_data['Spending Score (1-100)'].std()
            
            if 'Gender' in cluster_data.columns:
                profile['gender_dist'] = cluster_data['Gender'].value_counts().to_dict()
            
            # Generate profile characterization
            if 'avg_income' in profile and 'avg_spending' in profile:
                avg_income = profile['avg_income']
                avg_spending = profile['avg_spending']
                
                if avg_income > 70 and avg_spending > 70:
                    profile['type'] = "๐Ÿ’Ž HIGH VALUE"
                    profile['description'] = "High income, high spending - Premium customers"
                elif avg_income > 70 and avg_spending < 40:
                    profile['type'] = "๐Ÿ’ผ CONSERVATIVE"
                    profile['description'] = "High income, low spending - Potential for upselling"
                elif avg_income < 40 and avg_spending > 70:
                    profile['type'] = "๐ŸŽฏ BUDGET SPENDERS"
                    profile['description'] = "Low income, high spending - Price-sensitive loyal customers"
                elif avg_income < 40 and avg_spending < 40:
                    profile['type'] = "๐Ÿ“‰ LOW ENGAGEMENT"
                    profile['description'] = "Low income, low spending - Need retention strategies"
                else:
                    profile['type'] = "โš–๏ธ BALANCED"
                    profile['description'] = "Moderate income and spending - Core customer base"
            
            profiles.append(profile)
        
        return profiles