Module: BehavioralAnalysis/AnomalyDetection

Anomaly Detection & Early Warning Intelligence Dashboard

Advanced statistical intelligence module implementing Z-score analysis for behavioral anomaly detection across Swedish Parliament activity (2002-2025). Provides real-time early warning capability for unusual patterns in voting, document production, and attendance metrics.

Intelligence Methodology

Z-Score Statistical Analysis:

• Detection Threshold: |Z| ≥ 2.0 (2 standard deviations)
• Severity Classification: CRITICAL (>3σ), HIGH (2-3σ), MODERATE (1-2σ), LOW (<1σ)
• Direction Detection: UNUSUALLY_HIGH, UNUSUALLY_LOW, WITHIN_NORMAL_RANGE
• Temporal Coverage: 23 years × 4 quarters = 92 time periods

Anomaly Categories

Three Primary Anomaly Types:

1. BALLOT_ANOMALY: Unusual voting patterns (frequency, participation, outcomes)
2. DOCUMENT_ANOMALY: Abnormal document production (motions, questions, bills)
3. ATTENDANCE_ANOMALY: Irregular attendance patterns (chamber, committee)

Early Warning System

Automated Alert Mechanism:

• CRITICAL Alerts: |Z| > 3.0, immediate notification (red banner)
• HIGH Alerts: |Z| > 2.5, elevated monitoring (orange banner)
• Alert Persistence: 24-hour dismissal period
• Alert History: Tracked in localStorage for pattern analysis

Data Pipeline Architecture

OSINT Data Sources:

• Primary: cia-data/seasonal/view_riksdagen_seasonal_anomaly_detection_sample.csv
• Fallback: CIA GitHub repository (authoritative source)
• Update Frequency: 1-hour cache with real-time monitoring

Data Validation:

• CSV schema validation (year, quarter, z_score, severity, type)
• Range validation (years: 2002-2025, quarters: 1-4, z_score: numeric)
• Missing data handling with graceful degradation

Visualization Intelligence

Chart.js Analytics (6 visualizations):

1. Timeline: Chronological anomaly progression (scatter plot)
2. Distribution: Z-score normal curve with outlier markers
3. Type Breakdown: Pie chart (Ballot vs Document anomalies)
4. Severity Heatmap: Year × Quarter grid with color intensity
5. Quarterly Trends: Bar chart (Q1-Q4 anomaly frequency)
6. Recent Anomalies: Table of last 5 critical/high anomalies

Analytical Use Cases

Intelligence Applications:

• Crisis Detection: Identify sudden activity spikes (scandals, emergencies)
• Electoral Cycles: Detect pre-election behavioral shifts
• Policy Deadlines: Monitor document submission anomalies
• Attendance Monitoring: Track participation irregularities
• Historical Comparison: Benchmark current vs. past patterns

GDPR & Privacy Compliance