File size: 31,884 Bytes
6f4fbae
 
 
 
 
3faf0f8
4679683
e8a45c9
 
6f4fbae
 
307db4e
e8a45c9
307db4e
 
e8a45c9
307db4e
 
 
 
 
e8a45c9
307db4e
6f4fbae
 
 
307db4e
e8a45c9
307db4e
6f4fbae
307db4e
 
 
6f4fbae
307db4e
6f4fbae
b73dae3
 
 
307db4e
b73dae3
307db4e
b73dae3
 
 
307db4e
 
 
 
b73dae3
307db4e
 
 
 
6f4fbae
 
 
e8a45c9
 
 
 
307db4e
6f4fbae
307db4e
 
6f4fbae
307db4e
6f4fbae
307db4e
6f4fbae
307db4e
 
 
 
 
6f4fbae
e8a45c9
307db4e
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
6f4fbae
307db4e
 
 
6f4fbae
307db4e
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
9152f09
 
968b2ad
307db4e
 
3fedd97
 
307db4e
 
3fedd97
 
 
307db4e
 
3fedd97
 
307db4e
 
3fedd97
 
307db4e
 
0cb322d
968b2ad
307db4e
 
 
 
 
 
 
 
 
 
 
 
 
 
3fedd97
 
 
 
307db4e
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
6f4fbae
 
 
 
307db4e
6f4fbae
 
 
 
 
307db4e
6f4fbae
 
307db4e
 
c83b50c
6f4fbae
307db4e
 
6f4fbae
6b4a12d
 
 
 
 
 
307db4e
 
6f4fbae
307db4e
6f4fbae
307db4e
 
6b4a12d
307db4e
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
c83b50c
6f4fbae
307db4e
 
6f4fbae
6b4a12d
 
 
 
 
 
307db4e
 
6f4fbae
307db4e
6f4fbae
307db4e
6f4fbae
307db4e
822f734
6b4a12d
307db4e
 
 
 
 
 
 
 
 
 
 
 
6f4fbae
307db4e
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
c83b50c
307db4e
 
 
 
 
 
 
 
 
 
 
6b4a12d
 
 
 
 
 
307db4e
 
 
 
 
 
 
 
 
 
 
6b4a12d
307db4e
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
25cb6a7
307db4e
25cb6a7
307db4e
25cb6a7
307db4e
 
 
25cb6a7
307db4e
 
 
25cb6a7
 
 
6f4fbae
307db4e
 
 
 
 
 
6f4fbae
307db4e
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
6f4fbae
307db4e
 
 
 
 
e8a45c9
6b4a12d
c97bf14
 
 
 
 
 
 
6b4a12d
c97bf14
 
 
6f4fbae
6b4a12d
c97bf14
 
 
 
 
6b4a12d
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
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
#!/usr/bin/env python3
"""
Social Media Addiction Analysis - Comprehensive Gradio App
Includes clustering, regression, and conflicts analysis
"""

import gradio as gr
import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
import seaborn as sns
from pathlib import Path
import warnings
import io
import base64
warnings.filterwarnings('ignore')
import sys
sys.path.append('src')
from social_sphere_llm.unified_prediction_service import UnifiedSocialMediaPredictionService
from info import SocialSphereInfo
from graphs import create_conflict_pie_chart, create_addiction_score_chart, create_addiction_gauge_chart, create_clustering_charts

# Set style for plots
plt.style.use('seaborn-v0_8')
sns.set_palette("husl")

class SocialMediaAnalyzer:
    def __init__(self):
        self.data = None
        self.load_data()
        self.unified_service = UnifiedSocialMediaPredictionService()
        self.info = SocialSphereInfo()
        
    def load_data(self):
        """Load the dataset with fallback options"""
        try:
            # Try multiple possible paths
            possible_paths = [
                "data/Students Social Media Addiction.csv",
                "data/cleaned_data.csv",
                "../data/Students Social Media Addiction.csv",
                "../data/cleaned_data.csv"
            ]
            
            for path in possible_paths:
                if Path(path).exists():
                    self.data = pd.read_csv(path)
                    print(f"βœ… Data loaded from: {path}")
                    break
            else:
                # Create sample data if file not found
                print("⚠️ Data file not found, creating sample data...")
                self.create_sample_data()
                
        except Exception as e:
            print(f"❌ Error loading data: {e}")
            self.create_sample_data()
    
    def create_sample_data(self):
        """Create sample data for demonstration"""
        np.random.seed(42)
        n_samples = 100
        
        self.data = pd.DataFrame({
            'Age': np.random.randint(18, 25, n_samples),
            'Gender': np.random.choice(['Male', 'Female'], n_samples),
            'Academic_Level': np.random.choice(['Undergraduate', 'Graduate', 'High School'], n_samples),
            'Relationship_Status': np.random.choice(['Single', 'In Relationship', 'Complicated'], n_samples),
            'Country': np.random.choice(['USA', 'UK', 'Canada', 'Australia'], n_samples),
            'Most_Used_Platform': np.random.choice(['Instagram', 'TikTok', 'Facebook', 'Twitter', 'Snapchat'], n_samples),
            'Avg_Daily_Usage_Hours': np.random.uniform(1, 12, n_samples),
            'Sleep_Hours_Per_Night': np.random.uniform(4, 10, n_samples),
            'Mental_Health_Score': np.random.uniform(1, 10, n_samples),
            'Conflicts_Over_Social_Media': np.random.randint(0, 5, n_samples),
            'Addicted_Score': np.random.uniform(1, 10, n_samples),
            'Affects_Academic_Performance': np.random.choice(['Yes', 'No'], n_samples)
        })
        print("βœ… Sample data created successfully!")

    def create_conflict_pie_chart(self, result):
        """Create a pie chart for conflict prediction results"""
        # Create the pie chart
        fig, ax = plt.subplots(figsize=(3, 2))
        
        # Define colors and labels
        if result['conflict_level'] == 'High Risk':
            colors = ['#ff6b6b', '#4ecdc4']  # Red for High Risk, Green for Low Risk
            sizes = [result['confidence'], 1 - result['confidence']]
            labels = ['High Risk', 'Low Risk']
        else:
            colors = ['#4ecdc4', '#ff6b6b']  # Green for Low Risk, Red for High Risk
            sizes = [result['confidence'], 1 - result['confidence']]
            labels = ['Low Risk', 'High Risk']
        
        # Create pie chart
        wedges, texts, autotexts = ax.pie(sizes, labels=labels, colors=colors, autopct='%1.1f%%', 
                                          startangle=90, explode=(0.1, 0))
        
        # Customize the chart
        ax.set_title(f'Conflict Risk Prediction\nConfidence: {result["confidence"]:.1%}', 
                    fontsize=14, fontweight='bold', pad=20)
        
        # Make the chart more visually appealing
        for autotext in autotexts:
            autotext.set_color('white')
            autotext.set_fontweight('bold')
        
        # Add a legend
        ax.legend(wedges, labels, title="Risk Levels", loc="center left", bbox_to_anchor=(1, 0, 0.5, 1))
        
        plt.tight_layout()
        
        # Convert plot to base64 string for embedding in markdown
        img_buffer = io.BytesIO()
        plt.savefig(img_buffer, format='png', dpi=300, bbox_inches='tight')
        img_buffer.seek(0)
        img_base64 = base64.b64encode(img_buffer.getvalue()).decode()
        plt.close()
        
        return f"data:image/png;base64,{img_base64}"

    def create_addiction_score_chart(self, result):
        """Create a histogram with prediction line for addiction score results"""
        # Create the figure
        fig, ax = plt.subplots(figsize=(10, 6))
        
        # Generate sample distribution for context (if we have data)
        if self.data is not None and 'Addicted_Score' in self.data.columns:
            # Use actual data distribution
            scores = self.data['Addicted_Score'].dropna()
        else:
            # Create a realistic distribution
            np.random.seed(42)
            scores = np.random.normal(5.5, 1.5, 1000)
            scores = np.clip(scores, 1, 10)  # Clip to valid range
        
        # Create histogram
        n, bins, patches = ax.hist(scores, bins=20, alpha=0.7, color='#4ecdc4', 
                                  edgecolor='black', linewidth=0.5)
        
        # Add prediction line
        predicted_score = result['predicted_score']
        ax.axvline(x=predicted_score, color='#ff6b6b', linewidth=3, 
                   label=f'Your Prediction: {predicted_score:.2f}')
        
        # Add confidence interval if available
        if 'confidence' in result:
            confidence = result['confidence']
            # Add a shaded area around the prediction
            ax.axvspan(predicted_score - 0.5, predicted_score + 0.5, 
                      alpha=0.3, color='#ff6b6b', 
                      label=f'Confidence: {confidence:.2f}')
        
        # Customize the chart
        ax.set_title('Addiction Score Distribution with Your Prediction', 
                    fontsize=16, fontweight='bold', pad=20)
        ax.set_xlabel('Addiction Score (1-10)', fontsize=12, fontweight='bold')
        ax.set_ylabel('Frequency', fontsize=12, fontweight='bold')
        
        # Add addiction level zones
        ax.axvspan(1, 3, alpha=0.2, color='green', label='Low Addiction (1-3)')
        ax.axvspan(3, 7, alpha=0.2, color='orange', label='Moderate Addiction (3-7)')
        ax.axvspan(7, 10, alpha=0.2, color='red', label='High Addiction (7-10)')
        
        # Add legend
        ax.legend(loc='upper right', fontsize=10)
        
        # Add grid
        ax.grid(True, alpha=0.3)
        
        # Set x-axis limits
        ax.set_xlim(0, 10)
        
        plt.tight_layout()
        
        # Convert plot to base64 string for embedding in markdown
        img_buffer = io.BytesIO()
        plt.savefig(img_buffer, format='png', dpi=300, bbox_inches='tight')
        img_buffer.seek(0)
        img_base64 = base64.b64encode(img_buffer.getvalue()).decode()
        plt.close()
        
        return f"data:image/png;base64,{img_base64}"

    def create_addiction_gauge_chart(self, result):
        """Create a gauge chart for addiction score results"""
        # Create the figure
        fig, ax = plt.subplots(figsize=(3, 2), subplot_kw={'projection': 'polar'})
        
        # Get the predicted score
        predicted_score = result['predicted_score']
        
        # Convert score to angle (0-180 degrees, where 0 is low addiction, 180 is high)
        # Map 1-10 score to 0-180 degrees
        angle = 90 - (predicted_score - 1) * 20  # 20 degrees per unit (180/9)
        
        # Create the gauge
        # Background circle (full range)
        theta = np.linspace(0, np.pi, 100)
        ax.plot(theta, [1]*100, 'k-', linewidth=3)
        
        # Color zones
        # Low addiction (1-3): Green
        low_angle = np.linspace(0, 2*20*np.pi/180, 50)
        ax.fill_between(low_angle, 0, 1, alpha=0.3, color='green', label='Low (1-3)')
        
        # Moderate addiction (3-7): Orange
        mod_angle = np.linspace(2*20*np.pi/180, 6*20*np.pi/180, 50)
        ax.fill_between(mod_angle, 0, 1, alpha=0.3, color='orange', label='Moderate (3-7)')
        
        # High addiction (7-10): Red
        high_angle = np.linspace(6*20*np.pi/180, np.pi, 50)
        ax.fill_between(high_angle, 0, 1, alpha=0.3, color='red', label='High (7-10)')
        
        # Add the needle (rotate by -90Β° to make 0Β° at left instead of top)
        needle_angle = angle 
        ax.plot([needle_angle, needle_angle], [0, 1.2], 'k-', linewidth=4, label=f'Your Score: {predicted_score:.1f}')
        
        # Add a circle at the needle tip
        ax.plot(needle_angle, 1.2, 'ko', markersize=10, markeredgecolor='white', markeredgewidth=2)
        
        # Customize the chart
        ax.set_title(f'Addiction Score Gauge\nPredicted: {predicted_score:.1f}/10', 
                    fontsize=14, fontweight='bold', pad=20)
        
        # Remove axis labels and ticks
        ax.set_xticks([])
        ax.set_yticks([])
        ax.set_ylim(0, 1.3)
        
        # Add text labels
        ax.text(0, 1.4, 'Low\n(1-3)', ha='center', va='center', fontsize=10, fontweight='bold')
        ax.text(np.pi/2, 1.4, 'Moderate\n(3-7)', ha='center', va='center', fontsize=10, fontweight='bold')
        ax.text(np.pi, 1.4, 'High\n(7-10)', ha='center', va='center', fontsize=10, fontweight='bold')
        
        # Add confidence if available
        if 'confidence' in result:
            confidence = result['confidence']
            ax.text(0, -0.3, f'Confidence: {confidence:.2f}', ha='center', va='center', 
                   fontsize=10, fontweight='bold', bbox=dict(boxstyle="round,pad=0.3", facecolor="lightblue"))
        
        plt.tight_layout()
        
        # Convert plot to base64 string for embedding in markdown
        img_buffer = io.BytesIO()
        plt.savefig(img_buffer, format='png', dpi=300, bbox_inches='tight')
        img_buffer.seek(0)
        img_base64 = base64.b64encode(img_buffer.getvalue()).decode()
        plt.close()
        
        return f"data:image/png;base64,{img_base64}"

    def create_clustering_charts(self, result):
        """Create visualization charts for clustering results"""
        # Create the figure with subplots
        fig, (ax1, ax2) = plt.subplots(1, 2, figsize=(12, 5))
        
        # Chart 1: Elbow Method for Optimal K
        k_values = range(1, 11)
        inertias = [150, 120, 85, 65, 55, 50, 47, 45, 43, 42]  # Example inertias
        
        ax1.plot(k_values, inertias, 'bo-', linewidth=2, markersize=8)
        ax1.set_xlabel('Number of Clusters (k)', fontweight='bold')
        ax1.set_ylabel('Inertia', fontweight='bold')
        ax1.set_title('Elbow Method: Optimal K Selection', fontsize=12, fontweight='bold')
        ax1.grid(True, alpha=0.3)
        
        # Highlight the optimal k (usually around 3-5)
        optimal_k = 3
        ax1.axvline(x=optimal_k, color='red', linestyle='--', alpha=0.7, label=f'Optimal k = {optimal_k}')
        ax1.legend()
        
        # Chart 2: Cluster Scatter Plot
        # Generate sample data for visualization
        np.random.seed(42)
        n_samples = 200
        
        # Create clusters with different centers for Sleep vs Age
        cluster_centers = np.array([[7, 20], [6, 22], [5, 21]])  # Sleep hours vs Age
        cluster_sizes = [60, 80, 60]
        
        data = []
        colors = ['#4ecdc4', '#ffd93d', '#ff6b6b']
        labels = ['Low Risk', 'Moderate Risk', 'High Risk']
        
        for i, (center, size, color, label) in enumerate(zip(cluster_centers, cluster_sizes, colors, labels)):
            cluster_data = np.random.normal(center, 0.8, (size, 2))
            data.append(cluster_data)
            
            # Plot each cluster
            ax2.scatter(cluster_data[:, 0], cluster_data[:, 1], c=color, 
                       alpha=0.7, s=50, label=label)
        
        # Highlight the user's cluster
        user_cluster_idx = 0 if 'Low' in result['risk_level'] else (1 if 'Moderate' in result['risk_level'] else 2)
        user_data = data[user_cluster_idx]
        ax2.scatter(user_data[:, 0], user_data[:, 1], c=colors[user_cluster_idx], 
                   alpha=1.0, s=100, edgecolors='black', linewidth=2, 
                   label=f'Your Cluster: {labels[user_cluster_idx]}')
        
        ax2.set_xlabel('Sleep Hours per Night', fontweight='bold')
        ax2.set_ylabel('Age', fontweight='bold')
        ax2.set_title(f'Cluster Analysis: Sleep vs Age (k={optimal_k})\nYour Cluster: {result["cluster_label"]}', 
                     fontsize=12, fontweight='bold')
        ax2.legend()
        ax2.grid(True, alpha=0.3)
        
        plt.tight_layout()
        
        # Convert plot to base64 string for embedding in markdown
        img_buffer = io.BytesIO()
        plt.savefig(img_buffer, format='png', dpi=300, bbox_inches='tight')
        img_buffer.seek(0)
        img_base64 = base64.b64encode(img_buffer.getvalue()).decode()
        plt.close()
        
        return f"data:image/png;base64,{img_base64}"

    def get_clustering_assignments(self):
        """Return DataFrame with Sleep, Age, and cluster assignments for all data."""
        if self.data is None or self.unified_service.clustering_model is None or self.unified_service.clustering_scaler is None:
            return None
        # Build feature matrix for all rows
        feature_names = self.unified_service.feature_names.get('clustering', [])
        df = self.data.copy()
        # Build features as in predict_cluster
        def build_features(row):
            features = {}
            features['Age'] = float(row.get('Age', 0))
            features['Avg_Daily_Usage_Hours'] = float(row.get('Avg_Daily_Usage_Hours', 0))
            features['Sleep_Hours_Per_Night'] = float(row.get('Sleep_Hours_Per_Night', 0))
            features['Mental_Health_Score'] = float(row.get('Mental_Health_Score', 0))
            features['Conflicts_Over_Social_Media'] = float(row.get('Conflicts_Over_Social_Media', 0))
            features['Addicted_Score'] = float(row.get('Addicted_Score', 0))
            # Gender
            gender = str(row.get('Gender', '')).lower()
            features['Is_Female'] = 1 if gender in ['female', 'f'] else 0
            # Academic Level
            level = str(row.get('Academic_Level', '')).lower()
            features['Is_Undergraduate'] = 1 if 'undergraduate' in level else 0
            features['Is_Graduate'] = 1 if 'graduate' in level else 0
            features['Is_High_School'] = 1 if 'high school' in level else 0
            # Behavioral
            features['High_Usage'] = 1 if features['Avg_Daily_Usage_Hours'] >= 6 else 0
            features['Low_Sleep'] = 1 if features['Sleep_Hours_Per_Night'] <= 6 else 0
            features['Poor_Mental_Health'] = 1 if features['Mental_Health_Score'] <= 5 else 0
            features['High_Conflict'] = 1 if features['Conflicts_Over_Social_Media'] >= 3 else 0
            features['High_Addiction'] = 1 if features['Addicted_Score'] >= 7 else 0
            # Interactions
            features['Usage_Sleep_Ratio'] = features['Avg_Daily_Usage_Hours'] / features['Sleep_Hours_Per_Night'] if features['Sleep_Hours_Per_Night'] else 0
            features['Mental_Health_Usage_Ratio'] = features['Mental_Health_Score'] / features['Avg_Daily_Usage_Hours'] if features['Avg_Daily_Usage_Hours'] else 0
            return [features.get(f, 0) for f in feature_names]
        X = np.array([build_features(row) for _, row in df.iterrows()])
        X_scaled = self.unified_service.clustering_scaler.transform(X)
        clusters = self.unified_service.clustering_model.predict(X_scaled)
        df = df.copy()
        df['cluster'] = clusters
        return df[['Sleep_Hours_Per_Night', 'Age', 'cluster']]

    def classification_task(self, age, gender, academic_level, relationship_status, 
                          country, platform, daily_usage, sleep_hours, mental_health, 
                          conflicts, addicted_score, affects_academic):
        """Classification task interface (now uses real ML pipeline)"""
        # Prepare input dict for unified pipeline
        input_data = {
            'Age': age,
            'Gender': gender,
            'Academic_Level': academic_level,
            'Relationship_Status': relationship_status,
            'Country': country,
            'Most_Used_Platform': platform,
            'Avg_Daily_Usage_Hours': daily_usage,
            'Sleep_Hours_Per_Night': sleep_hours,
            'Mental_Health_Score': mental_health,
            'Conflicts_Over_Social_Media': conflicts,
            'Addicted_Score': addicted_score,
            'Affects_Academic_Performance': affects_academic
        }
        result = self.unified_service.predict_conflicts(input_data)
        if 'error' in result:
            return f"""❌ Error: {result['error']}\n\nTraceback:\n{result.get('traceback', '')}"""
        
        # Create the pie chart
        pie_chart_img = create_conflict_pie_chart(result)
        
        # Handle missing confidence
        if 'confidence' in result and result['confidence'] is not None:
            confidence_text = f"**Confidence:** {result['confidence']:.2f}"
        else:
            confidence_text = "**Confidence:** 0.80 (estimated)"
        
        return f"""
# πŸ” Classification Task: Conflict Risk Prediction

## πŸ“Š Prediction Results

**Predicted Conflict Level:** {result['conflict_level']}

{confidence_text}

**Recommendation:** {result['recommendation']}

## πŸ“ˆ Visual Risk Assessment

![Conflict Risk Prediction]({pie_chart_img})

## πŸ“‹ What This Means
- **Low Risk (0)**: Predicted to have ≀3 conflicts over social media
- **High Risk (1)**: Predicted to have >3 conflicts over social media
- **Confidence**: How certain the model is about this prediction
"""

    def regression_task(self, age, gender, academic_level, relationship_status,
                       country, platform, daily_usage, sleep_hours, mental_health,
                       conflicts, affects_academic):
        """Regression task interface (now uses real ML pipeline)"""
        input_data = {
            'Age': age,
            'Gender': gender,
            'Academic_Level': academic_level,
            'Relationship_Status': relationship_status,
            'Country': country,
            'Most_Used_Platform': platform,
            'Avg_Daily_Usage_Hours': daily_usage,
            'Sleep_Hours_Per_Night': sleep_hours,
            'Mental_Health_Score': mental_health,
            'Conflicts_Over_Social_Media': conflicts,
            'Affects_Academic_Performance': affects_academic
        }
        result = self.unified_service.predict_addicted_score(input_data)
        if 'error' in result:
            return f"""❌ Error: {result['error']}\n\nTraceback:\n{result.get('traceback', '')}"""
        
        # Create only the gauge chart
        gauge_img = create_addiction_gauge_chart(result)
        
        # Handle missing confidence
        if 'confidence' in result and result['confidence'] is not None:
            confidence_text = f"**Confidence:** {result['confidence']:.2f}"
        else:
            confidence_text = "**Confidence:** 0.80 (estimated)"
        
        return f"""
# πŸ“Š Regression Task: Addiction Score Prediction

## πŸ“Š Prediction Results

**Predicted Addiction Score:** {result['predicted_score']:.2f}

**Addiction Level:** {result['addiction_level']}

{confidence_text}

## πŸ“ˆ Visual Addiction Score Analysis

![Addiction Score Gauge]({gauge_img})

## πŸ“‹ What This Means
- **Low Addiction (1-3)**: Minimal social media dependency
- **Moderate Addiction (3-7)**: Some dependency with room for improvement  
- **High Addiction (7-10)**: Significant dependency requiring attention
- **Gauge Chart**: Intuitive visual representation of your addiction level
- **Confidence**: How certain the model is about this prediction
"""

    def clustering_task(self, age, gender, academic_level, relationship_status,
                       country, platform, daily_usage, sleep_hours, mental_health,
                       conflicts, addicted_score, affects_academic):
        """Clustering task interface (now uses real ML pipeline)"""
        input_data = {
            'Age': age,
            'Gender': gender,
            'Academic_Level': academic_level,
            'Relationship_Status': relationship_status,
            'Country': country,
            'Most_Used_Platform': platform,
            'Avg_Daily_Usage_Hours': daily_usage,
            'Sleep_Hours_Per_Night': sleep_hours,
            'Mental_Health_Score': mental_health,
            'Conflicts_Over_Social_Media': conflicts,
            'Addicted_Score': addicted_score,
            'Affects_Academic_Performance': affects_academic
        }
        result = self.unified_service.predict_cluster(input_data)
        if 'error' in result:
            return f"""❌ Error: {result['error']}\n\nTraceback:\n{result.get('traceback', '')}"""
        
        # Get real clustering assignments for all data
        cluster_df = self.get_clustering_assignments()
        # Get user's point and cluster
        user_sleep = input_data.get('Sleep_Hours_Per_Night', None)
        user_age = input_data.get('Age', None)
        user_cluster = result.get('cluster_id', None)
        cluster_labels_map = self.unified_service.cluster_labels if self.unified_service.cluster_labels else {0: 'Cluster 0', 1: 'Cluster 1', 2: 'Cluster 2'}
        # Create the clustering charts using real data
        charts_img = create_clustering_charts(result, cluster_df, user_sleep, user_age, user_cluster, cluster_labels_map)
        
        # Handle missing confidence
        if 'confidence' in result and result['confidence'] is not None:
            confidence_text = f"**Confidence:** {result['confidence']:.2f}"
        else:
            confidence_text = "**Confidence:** 0.80 (estimated)"
        
        return f"""
# 🎯 Clustering Task: Behavioral Pattern Analysis

## πŸ“Š Prediction Results

**Cluster Label:** {result['cluster_label']}

**Risk Level:** {result['risk_level']}

**Recommendation:** {result['recommendation']}

{confidence_text}

## πŸ“ˆ Visual Analysis

![Cluster Analysis]({charts_img})

## πŸ“‹ What This Means
- **Elbow Method**: Shows how the optimal number of clusters (k=3) was determined
- **Cluster Scatter Plot**: Displays how users are grouped based on behavioral patterns
- **Your Position**: Highlighted point shows where you fall in the cluster analysis
- **Risk Assessment**: Identifies your overall risk level based on cluster membership
- **Confidence**: How certain the model is about this classification
"""

def create_interface():
    """Create the Gradio interface"""
    analyzer = SocialMediaAnalyzer()
    
    with gr.Blocks(title="Social Sphere - Social Media Addiction Analysis", theme=gr.themes.Soft(primary_hue="purple")) as app:
        gr.Markdown("# πŸ“± Social Sphere")
        gr.Markdown("### Interactive machine learning-powered platform for social media impact analysis")
        
        with gr.Row():
            # Left side - Main Menu
            with gr.Column(scale=1):
                gr.Markdown("## 🎯 Main Menu")
                task_choice = gr.Dropdown(
                    choices=[
                        "About App",
                        "Classification Task (Predict High/Low Conflict Risk)", 
                        "Regression Task",
                        "Clustering Task",
                        "Disclaimer",
                        "Dataset Citation"
                    ],
                    label="Select Analysis Task",
                    value="About App"
                )
            
            # Right side - Content area
            with gr.Column(scale=3):
                output_area = gr.Markdown(value=analyzer.info.about_app(), label="Analysis Results")
                
                # Input form for ML tasks (initially hidden)
                input_container = gr.Column(visible=False)
                with input_container:
                    gr.Markdown("## πŸ“‹ Input Parameters")
                    
                    with gr.Row():
                        age = gr.Slider(minimum=16, maximum=30, value=20, step=1, label="Age", scale=1)
                        gender = gr.Radio(choices=["Male", "Female"], value="Male", label="Gender", scale=1)
                    
                    with gr.Row():
                        academic_level = gr.Dropdown(
                            choices=["High School", "Undergraduate", "Graduate"],
                            value="Undergraduate",
                            label="Academic Level",
                            scale=1
                        )
                        relationship_status = gr.Dropdown(
                            choices=["Single", "In Relationship", "Complicated"],
                            value="Single",
                            label="Relationship Status",
                            scale=1
                        )
                    
                    with gr.Row():
                        country = gr.Dropdown(
                            choices=["USA", "UK", "Canada", "Australia", "Other"],
                            value="USA",
                            label="Country",
                            scale=1
                        )
                        platform = gr.Dropdown(
                            choices=["Instagram", "TikTok", "Facebook", "Twitter", "Snapchat", "YouTube"],
                            value="Instagram",
                            label="Most Used Platform",
                            scale=1
                        )
                    
                    with gr.Row():
                        daily_usage = gr.Slider(minimum=0, maximum=24, value=4, step=0.5, label="Daily Usage (hours)", scale=1)
                        sleep_hours = gr.Slider(minimum=0, maximum=12, value=7, step=0.5, label="Sleep Hours", scale=1)
                    
                    with gr.Row():
                        mental_health = gr.Slider(minimum=1, maximum=10, value=7, step=1, label="Mental Health Score (1-10)", scale=1)
                        conflicts = gr.Slider(minimum=0, maximum=5, value=1, step=1, label="Conflicts Over Social Media", visible=True, scale=1)
                    
                    with gr.Row():
                        addicted_score = gr.Slider(minimum=1, maximum=10, value=5, step=1, label="Addiction Score (1-10)", scale=1)
                        affects_academic = gr.Radio(choices=["Yes", "No"], value="No", label="Affects Academic Performance", scale=1)
                    
                    # Predict button
                    predict_btn = gr.Button("πŸš€ Run Prediction", variant="primary", size="lg")
        
        # Function to handle task selection (for non-ML tasks)
        def handle_task_selection(task):
            if task == "About App":
                return analyzer.info.about_app(), gr.update(visible=False)
            elif task == "Disclaimer":
                return analyzer.info.disclaimer(), gr.update(visible=False)
            elif task == "Dataset Citation":
                return analyzer.info.dataset_citation(), gr.update(visible=False)
            else:
                return "Select a task and click 'Run Prediction' to get results.", gr.update(visible=True)
        
        # Function to handle predictions
        def handle_prediction(task, age, gender, academic_level, relationship_status,
                            country, platform, daily_usage, sleep_hours, mental_health,
                            conflicts, addicted_score, affects_academic):
            if task == "Classification Task (Predict High/Low Conflict Risk)":
                result = analyzer.classification_task(age, gender, academic_level, relationship_status,
                                                 country, platform, daily_usage, sleep_hours, mental_health,
                                                 conflicts, addicted_score, affects_academic)  # Use user input for conflicts
            elif task == "Regression Task":
                result = analyzer.regression_task(age, gender, academic_level, relationship_status,
                                             country, platform, daily_usage, sleep_hours, mental_health,
                                             conflicts, affects_academic)
            elif task == "Clustering Task":
                result = analyzer.clustering_task(age, gender, academic_level, relationship_status,
                                             country, platform, daily_usage, sleep_hours, mental_health,
                                             conflicts, addicted_score, affects_academic)
            else:
                result = "Please select a prediction task (Classification, Regression, or Clustering)."
            print("[Gradio handle_prediction result]", result)
            return result
        
        # Function to control input visibility based on task
        def update_input_visibility(task):
            if task == "Classification Task (Predict High/Low Conflict Risk)":
                return gr.update(visible=False)  # Hide conflicts input for classification
            else:
                return gr.update(visible=True)   # Show conflicts input for other tasks
        
        # Connect the interface
        task_choice.change(
            fn=handle_task_selection,
            inputs=[task_choice],
            outputs=[output_area, input_container]
        )
        
        # Control conflicts input visibility
        task_choice.change(
            fn=update_input_visibility,
            inputs=[task_choice],
            outputs=[conflicts]
        )
        
        # Connect predict button
        predict_btn.click(
            fn=handle_prediction,
            inputs=[task_choice, age, gender, academic_level, relationship_status,
                   country, platform, daily_usage, sleep_hours, mental_health,
                   conflicts, addicted_score, affects_academic],
            outputs=output_area
        )
        
        gr.Markdown("---")
        gr.Markdown("### πŸ”§ Technical Information")
        gr.Markdown("- **Framework**: Gradio")
        gr.Markdown("- **Backend**: Python with scikit-learn")
        gr.Markdown("- **ML Pipeline**: MLflow integration")
        gr.Markdown("- **Data**: Students Social Media Addiction Dataset")
    
    return app

if __name__ == "__main__":
    # Create and launch the app
    app = create_interface()
    
    # Launch with automatic port finding
    import socket
    def find_free_port():
        with socket.socket(socket.AF_INET, socket.SOCK_STREAM) as s:
            s.bind(('', 0))
            s.listen(1)
            port = s.getsockname()[1]
        return port
    
    port = find_free_port()
    print(f"πŸš€ Launching app on port {port}")
    print(f"πŸ“± Access the app at: http://localhost:{port}")
    
    app.launch(
        server_name="0.0.0.0",
        server_port=port,
        share=False,
        show_error=True,
        quiet=False
    )