🏛️ Comprehensive Political Data Architecture for Democratic Transparency
🗄️ 50+ Years Historical Data · 15 CIA Data Subsystems · 14-Language Support

📋 Document Owner: CEO | 📄 Version: 1.2 | 📅 Last Updated: 2026-04-20 (UTC)
🔄 Review Cycle: Annual | ⏰ Next Review: 2027-04-20
🏢 Owner: Hack23 AB (Org.nr 5595347807) | 🏷️ Classification: Public

🆕 What changed since last review (v1.1 → v1.2, 2026-04-20):

• 📈 IMF Open Data added as a primary external data source (§4.1.5 below) alongside SCB and World Bank, per ADR 0001 (accepted 2026-04-20). Economic Data Contract v2.0 (effective 2026-04-20; v1 grace → 2026-05-31) extends the economic data-point schema with additive fields: source.imf[], dataPoints[].provider, dataPoints[].projection (boolean), dataPoints[].projectionVintage (e.g., "WEO-2026-04"). Pure-TypeScript client scripts/imf-client.ts (no MCP — SBOM-covered via npm). Cache: analysis/data/imf/{indicator}/{country}.json + .meta.json.

🆕 What changed since last review (v1.0 → v1.1, 2026-04-20):

• Factual correction: the cia-data/ tree currently materialises 15 subsystems (not 19 "products"): anomaly, coalition, committee, distribution, election, election-cycle, ministry, parties, party, percentile, politician, pre-election, risk, seasonal, voting. All headline counts, ToC entries, and inventory tables have been reconciled with the filesystem.
• Added documentation of the npm-package data contract: typed subpath exports ./cia/*, ./dashboards/*, ./shared/*, ./ui/* in package.json expose TypeScript .d.ts surfaces generated from schemas/ via generate-types-from-cia-schemas.
• Documented schema governance scripts under scripts/: sync-cia-schemas, validate-against-cia-schemas, check-cia-schema-updates, generate-types-from-cia-schemas (all Node 25 native TypeScript modules).
• Added news-corpus footprint: 2,669 files under news/ across 14 languages (EN, SV, DA, NB, DE, ES, FI, FR, HE RTL, AR RTL, JA, KO, NL, ZH). News metadata extracted by extract-news-metadata feeds generate-rss, generate-sitemap, generate-sitemap-html, and generate-news-indexes.
• Validation stack locked: ajv 8.18.0 (JSON Schema), happy-dom 20.9.0 (DOM test), Vitest 4.1.4 (unit), Cypress 15.14.0 (E2E). All validations invoked in prebuild / CI.
• Aligned with Secure_Development_Policy §7 "Data Integrity", CLASSIFICATION, ISO 27001:2022 A.5.33/A.8.12, NIST CSF 2.0 PR.DS-1/PR.DS-6/ID.AM-3, CIS Controls v8.1 #3, GDPR Art. 5/32.

This document defines the data model for the Riksdagsmonitor platform, documenting entity relationships, data structures, CIA product schemas, and data quality metrics for Swedish Parliament political data spanning 50+ years.

Riksdagsmonitor maintains a comprehensive data architecture integrating 50+ years of Swedish Parliament data (1971-2026) with 15 data subsystems from the CIA platform, surfaced through the cia-data/ tree in this repository and re-exported as typed subpaths (./cia/*, ./dashboards/*, ./shared/*, ./ui/*) in the public riksdagsmonitor npm package (SLSA provenance attested). This document defines all data entities, relationships, schemas, pipelines, and integration patterns following Hack23 AB's ISMS standards (ISO 27001:2022, NIST CSF 2.0, CIS Controls v8.1, GDPR, NIS2).

Key Statistics:

• 2,494 Politicians (349 current MPs)
• 3.5M+ Voting Records across all parliaments
• 109,000+ Documents (motions, propositions, reports)
• 8 Political Parties + 40 historical parties
• 15 Committees with complete assignment tracking
• 20 Governments with 76 roles and 500 role members
• 14 Languages with full multi-language support
• 15 CIA Data Subsystems materialised under cia-data/ (anomaly, coalition, committee, distribution, election, election-cycle, ministry, parties, party, percentile, politician, pre-election, risk, seasonal, voting) with 50+ CSV data files

1. Political Entities & Data Dictionary
2. CIA Data Subsystems (15 Subsystems)
3. Entity-Relationship Diagrams
4. Data Sources
5. Data Schemas & Validation
6. Data Pipeline Architecture
7. Multi-Language Data Architecture
8. Performance & Caching
9. C4 Model Integration
10. ISMS Compliance

Table: person_data
Records: 2,494 (349 active MPs)
Source: Swedish Riksdag API + CIA Platform
Update Frequency: Daily

Indexes:

• Primary Key: person_id
• Foreign Key: party → sweden_political_party.party_id
• Index: status, party, risk_level

Business Rules:

• Active MPs: status = 'Tjänstgörande riksdagsledamot'
• Risk threshold: HIGH when risk_score >= 50
• Historical coverage: 1971-2026

Table: sweden_political_party
Records: 40 (12 riksdag parties, 28 historical)
Source: Swedish Riksdag API + Election Authority
Update Frequency: Monthly (on party changes)

Indexes:

• Primary Key: party_id
• Index: riksdag_status, founded_year

Business Rules:

• Active Riksdag Parties (8): S, M, SD, C, V, MP, KD, L
• Historical Parties: ny, pp, v(k), fp, etc. (32 parties)
• Threshold: Riksdag representation requires >= 4% vote share

Party Descriptions:

Table: committee_document_data
Records: 8,740 committee documents
Source: Swedish Riksdag API
Update Frequency: Daily

Indexes:

• Primary Key: committee_id + document_id
• Foreign Key: document_id → document_data.document_id
• Index: published_date, document_type

Swedish Riksdag Committees (15):

Table: document_data
Records: 109,259
Source: Swedish Riksdag API
Update Frequency: Daily

Indexes:

• Primary Key: document_id
• Index: document_type, rm, published_date, organ
• Full-text index: title, subtitle, fulltext

Document Types:

Table: vote_data
Records: 3,529,786
Source: Swedish Riksdag API
Update Frequency: Real-time (after votes)

Indexes:

• Primary Key: vote_id
• Foreign Keys: person_id, party, ballot_id, document_id
• Index: vote_date, party, is_rebel_vote

Vote Classifications:

• Ja (Yes): Approval vote
• Nej (No): Rejection vote
• Avstår (Abstain): Abstention
• Frånvarande (Absent): Not present

Metrics Derived:

• Win Rate: (winning_votes / total_votes) * 100
• Rebel Rate: (rebel_votes / total_votes) * 100
• Attendance Rate: ((total_votes - absent) / total_ballots) * 100

Table: government_body_data
Records: 6,520 (20 governments, 76 roles, 500 role members)
Source: Swedish Government + CIA Platform
Update Frequency: On government changes

Indexes:

• Primary Key: ministry_id
• Foreign Keys: government_id, minister_person_id, party
• Index: start_date, end_date, effectiveness_score

Current Swedish Ministries (11):

Table: government_role_data
Records: 76 roles, 500 role members
Description: Cabinet positions and government appointments

Role Types:

• Minister (Minister): Cabinet minister
• Statssekreterare (State Secretary): Senior civil servant
• Politiskt sakkunnig (Political Advisor): Political staff
• Pressekreterare (Press Secretary): Communications

Note (v1.1, 2026-04-20): The prior framing of "19 Products" has been retired. The current repository materialises 15 subsystems under cia-data/, each backed by one or more CSV extracts and (where applicable) a JSON Schema under schemas/. The subsystem list below is the canonical, filesystem-verified inventory.

Product Name: Riksdag Intelligence Overview
Purpose: Comprehensive snapshot of parliamentary activity
Data Files:

• view_riksdagen_politician_sample.csv
• view_riksdagen_party_summary_sample.csv
• cia-data/production-stats.json

Key Fields & Metrics:

• Total active MPs (349)
• Total votes cast (3.5M+)
• Total documents (109K+)
• Party representation breakdown
• Committee assignments
• Risk score distribution

Dashboard Integration: Homepage overview section
Update Frequency: Daily (03:00 CET)

Product Name: Party Performance & Effectiveness Analytics
Purpose: Longitudinal party analysis (1990-2026, 37 years)
Data Files:

• cia-data/party/view_party_effectiveness_trends_sample.csv
• cia-data/party/view_party_performance_metrics_sample.csv
• cia-data/party/distribution_party_effectiveness_trends.csv
• cia-data/party/distribution_party_momentum.csv

Key Fields & Metrics:

Dashboard Integration: Party Performance Dashboard (index.html)
Visualizations:

• Effectiveness trends timeline (Chart.js)
• Party comparison scatter plot (D3.js)
• Momentum indicator gauges
• Coalition alignment matrix

Update Frequency: Daily

Product Name: Ministry Performance Scorecards
Purpose: Cabinet-level analysis and ministry effectiveness
Data Files:

• cia-data/ministry/distribution_ministry_effectiveness.csv
• cia-data/ministry/distribution_ministry_decision_impact.csv
• cia-data/ministry/distribution_ministry_productivity_matrix.csv
• cia-data/ministry/distribution_ministry_risk_levels.csv

Key Fields & Metrics:

Dashboard Integration: Ministry Dashboard (placeholder)
Visualizations:

• Ministry effectiveness radar chart
• Decision impact heatmap (D3.js)
• Risk level gauge chart
• Productivity matrix scatter plot

Update Frequency: Weekly

Product Name: Historical Patterns & Trend Forecasting
Purpose: Election cycle intelligence (1994-2034, 9 cycles)
Data Files:

• cia-data/election-cycle/view_election_cycle_comparative_analysis_sample.csv (1,110 records, 153KB)
• cia-data/election-cycle/view_election_cycle_decision_intelligence_sample.csv (414 records, 59KB)
• cia-data/election-cycle/view_election_cycle_predictive_intelligence_sample.csv (41 records, 3.9KB)
• cia-data/election-cycle/view_election_cycle_temporal_trends_sample.csv (74 records, 8.7KB)

Key Fields & Metrics:

Dashboard Integration: Election Cycle Dashboard (js/election-cycle-dashboard.js)
Visualizations:

• Multi-cycle performance timeline (Chart.js)
• Party tier distribution (D3.js)
• Risk forecast scatter chart
• Temporal trends multi-axis chart

Update Frequency: Daily

Product Name: Politician Influence Network Analysis
Purpose: Identify MPs with highest influence scores
Data Files:

• cia-data/politician/view_riksdagen_politician_influence_metrics_sample.csv

Key Fields:

• person_id, first_name, last_name, party
• influence_score (0-100)
• network_centrality (0-1)
• committee_influence
• cross_party_connections

Ranking Criteria: Influence score (descending)
Dashboard Integration: Politician Dashboard
Update Frequency: Weekly

Product Name: Legislative Output Analysis
Purpose: Rank MPs by document production
Data Files:

• cia-data/politician/view_riksdagen_politician_sample.csv

Key Fields:

• person_id, first_name, last_name, party
• documents_last_year (INTEGER)
• total_documents (INTEGER)
• document_types (JSON array)

Ranking Criteria: Documents produced (last 12 months, descending)
Dashboard Integration: Politician Dashboard
Update Frequency: Daily

Product Name: Voting Pattern Outlier Detection
Purpose: Identify MPs with highest rebellion rates
Data Files:

• cia-data/politician/view_politician_behavioral_trends_sample.csv

Key Fields:

• person_id, first_name, last_name, party
• annual_rebel_rate (0-100%)
• controversial_votes (count)
• party_discipline_deviation

Ranking Criteria: Rebel rate (descending)
Dashboard Integration: Politician Dashboard
Update Frequency: Daily

Product Name: Attendance Tracking
Purpose: Monitor MP absences
Data Files:

• cia-data/politician/view_politician_risk_summary_sample.csv

Key Fields:

• person_id, first_name, last_name, party
• annual_absence_rate (0-100%)
• absenteeism_violations (count)
• attendance_trend

Ranking Criteria: Absence rate (descending)
Dashboard Integration: Politician Dashboard
Update Frequency: Daily

Product Name: Cross-Party Voting Analysis
Purpose: Identify MPs who frequently vote against party line
Data Files:

• cia-data/voting/distribution_voting_anomaly_classification.csv

Key Fields:

• person_id, party, rebel_vote_count
• party_line_deviation_pct
• coalition_alignment_score

Ranking Criteria: Rebel votes (descending)
Dashboard Integration: Politician Dashboard
Update Frequency: Daily

Product Name: Cross-Party Collaboration Patterns
Purpose: Identify MPs facilitating coalition work
Data Files:

• cia-data/coalition/distribution_coalition_alignment.csv

Key Fields:

• person_id, coalition_score
• cross_party_proposals
• bridge_connections

Ranking Criteria: Coalition score (descending)
Dashboard Integration: Coalition Dashboard
Update Frequency: Weekly

Product Name: Emerging Political Figures
Purpose: Identify rapidly advancing politicians
Data Files:

• cia-data/politician/view_politician_behavioral_trends_sample.csv

Key Fields:

• person_id, career_trajectory
• momentum_score
• media_mentions_growth
• influence_acceleration

Ranking Criteria: Momentum score (descending)
Dashboard Integration: Politician Dashboard
Update Frequency: Weekly

Product Name: MPs at Election Risk
Purpose: Predict electoral vulnerability
Data Files:

• cia-data/politician/view_politician_risk_summary_sample.csv

Key Fields:

• person_id, risk_score (0-100)
• risk_level (LOW, MEDIUM, HIGH, CRITICAL)
• risk_factors (JSON array)

Ranking Criteria: Risk score (descending)
Dashboard Integration: Politician Dashboard
Update Frequency: Daily

Product Name: Rule Violation Tracking
Purpose: Monitor transparency and ethics violations
Data Files:

• cia-data/politician/view_politician_risk_summary_sample.csv

Key Fields:

• person_id, total_violations
• violation_types (effectiveness, discipline, productivity, collaboration)
• latest_violation_date

Ranking Criteria: Total violations (descending)
Dashboard Integration: Politician Dashboard
Update Frequency: Daily

Product Name: Public Visibility Index
Purpose: Track media mentions and public visibility
Data Files:

• External media analysis (placeholder)

Key Fields:

• person_id, media_mentions_count
• visibility_score (0-100)
• sentiment_average (-1 to +1)

Ranking Criteria: Media mentions (descending)
Dashboard Integration: Placeholder
Update Frequency: Weekly

Product Name: Committee Influence Mapping
Purpose: Visualize committee assignments and influence
Data Files:

• cia-data/committee/view_riksdagen_committee_decisions.csv
• cia-data/committee/distribution_annual_committee_documents.csv

Key Fields & Metrics:

• committee_id, committee_name
• assigned_members (JSON array)
• document_count, decision_count
• productivity_score (0-100)
• influence_centrality (0-1)

Dashboard Integration: Committee Dashboard
Visualizations:

• Network graph (D3.js force-directed)
• Assignment matrix heatmap
• Productivity comparison bar chart

Update Frequency: Weekly

Product Name: Career Trajectory Tracking
Purpose: Analyze politician career paths and milestones
Data Files:

• cia-data/politician/view_riksdagen_politician_experience_summary_sample.csv
• cia-data/politician/distribution_experience_levels.csv

Key Fields & Metrics:

• person_id, career_start_date, total_years
• roles_held (JSON array)
• committee_assignments_history
• government_positions
• experience_level (JUNIOR, INTERMEDIATE, SENIOR, VETERAN)

Dashboard Integration: Politician Dashboard
Visualizations:

• Career timeline (Gantt chart)
• Experience distribution (Chart.js)
• Role progression flowchart

Update Frequency: Monthly

Product Name: 50+ Years of Party Evolution
Purpose: Historical party performance (1971-2026)
Data Files:

• cia-data/party/view_party_effectiveness_trends_sample.csv
• cia-data/party/distribution_annual_party_votes.csv
• cia-data/party/distribution_annual_party_members.csv

Key Fields & Metrics:

• party, year, decade
• member_count, vote_count
• avg_win_rate, avg_discipline
• electoral_success_rate
• historical_trend (RISING, DECLINING, STABLE)

Dashboard Integration: Party Performance Dashboard
Visualizations:

• 50-year timeline (Chart.js)
• Party evolution heatmap (D3.js)
• Electoral cycle comparison

Update Frequency: Daily

Product Name: Quarterly Parliamentary Activity Analysis
Purpose: Identify seasonal trends and patterns (2002-2025)
Data Files:

• cia-data/seasonal/view_riksdagen_seasonal_activity_patterns_sample.csv (85 records)

Key Fields & Metrics:

• year, quarter, is_election_year, election_cycle
• total_ballots, active_politicians, attendance_rate
• documents_produced, decisions_made
• q_baseline_ballots, q_baseline_docs, q_baseline_attendance
• ballot_z_score, doc_z_score, attendance_z_score
• base_activity_classification (LOW, MODERATE, HIGH, VERY_HIGH)
• seasonal_pattern_classification
• cross_year_quarter_avg_ballots, cross_year_z_score
• qoq_ballot_change_pct, activity_quartile_cycle

Dashboard Integration: Seasonal Activity Patterns Dashboard
Visualizations:

• Quarterly heatmap (D3.js)
• Time series chart (Chart.js)
• Z-score distribution
• Cross-year comparison

Update Frequency: Quarterly

Product Name: Statistical Outlier Identification
Purpose: Real-time anomaly detection (2002-2026, 41 quarters)
Data Files:

• cia-data/seasonal/view_riksdagen_seasonal_anomaly_detection_sample.csv (41 records)

Key Fields & Metrics:

• year, quarter, is_election_year, parliamentary_period
• total_ballots, active_politicians, attendance_rate, documents_produced
• q_baseline_ballots, q_baseline_docs, q_baseline_attendance
• q_stddev_ballots, q_stddev_docs, q_stddev_attendance
• ballot_z_score, doc_z_score, attendance_z_score
• activity_classification (NORMAL, UNUSUALLY_LOW, UNUSUALLY_HIGH)
• anomaly_type (Ballot, Document, Attendance, Mixed)
• anomaly_direction (UNUSUALLY_HIGH, UNUSUALLY_LOW, NORMAL)
• max_z_score, anomaly_severity (LOW, MODERATE, HIGH, CRITICAL)
• quarter_label (Q1_JAN_MAR, Q2_APR_JUN, Q3_JUL_SEP, Q4_OCT_DEC)

Anomaly Detection Criteria:

• |Z| < 1.5: LOW severity (within normal range)
• 1.5 ≤ |Z| < 2.0: MODERATE severity
• 2.0 ≤ |Z| < 2.5: HIGH severity
• |Z| ≥ 2.5: CRITICAL severity

Historical Findings (41 quarters):

• 8 CRITICAL anomalies (Z ≥ 2.5)
• 2 HIGH anomalies (2.0 ≤ Z < 2.5)
• 12 MODERATE anomalies (1.5 ≤ Z < 2.0)
• 19 LOW/NORMAL (normal activity)
• Most extreme: 2006 Q1 document anomaly (Z = +10.97)

Dashboard Integration: Anomaly Detection Dashboard (index.html, inline script)
Visualizations:

• Anomaly timeline (Chart.js)
• Z-score distribution (histogram)
• Severity classification (pie chart)
• Heatmap (D3.js)
• Alert panel (critical anomalies)

Update Frequency: Quarterly

erDiagram
    POLITICIAN ||--o{ VOTE : casts
    POLITICIAN }o--|| PARTY : belongs_to
    POLITICIAN }o--o{ COMMITTEE : assigned_to
    POLITICIAN ||--o{ DOCUMENT : authors
    POLITICIAN }o--o{ GOVERNMENT_ROLE : holds
    
    PARTY ||--o{ POLITICIAN : has_members
    PARTY ||--o{ MINISTRY : controls
    
    DOCUMENT ||--o{ VOTE : triggers
    DOCUMENT }o--|| COMMITTEE : processed_by
    DOCUMENT }o--o{ DOCUMENT : references
    
    COMMITTEE ||--o{ DOCUMENT : produces
    COMMITTEE }o--o{ POLITICIAN : includes
    
    MINISTRY ||--o{ GOVERNMENT_ROLE : contains
    MINISTRY }o--|| PARTY : led_by
    MINISTRY ||--o{ DOCUMENT : issues
    
    GOVERNMENT_ROLE }o--|| POLITICIAN : held_by
    GOVERNMENT_ROLE }o--|| MINISTRY : part_of
    
    POLITICIAN {
        varchar person_id PK
        varchar first_name
        varchar last_name
        varchar party FK
        varchar status
        decimal risk_score
        varchar risk_level
    }
    
    PARTY {
        varchar party_id PK
        varchar party_name
        integer founded_year
        varchar riksdag_status
        decimal avg_win_rate
    }
    
    COMMITTEE {
        varchar committee_id PK
        varchar committee_name
        integer total_documents
        decimal productivity_score
    }
    
    DOCUMENT {
        varchar document_id PK
        varchar document_type
        date published_date
        varchar status
        varchar organ FK
    }
    
    VOTE {
        varchar vote_id PK
        varchar ballot_id FK
        varchar person_id FK
        varchar party FK
        varchar vote
        date vote_date
        boolean is_rebel_vote
    }
    
    MINISTRY {
        varchar ministry_id PK
        varchar ministry_name
        varchar minister_person_id FK
        varchar party FK
        decimal effectiveness_score
        varchar risk_level
    }
    
    GOVERNMENT_ROLE {
        varchar role_id PK
        varchar role_name
        varchar role_type
        varchar ministry_id FK
        varchar person_id FK
        date start_date
        date end_date
    }

erDiagram
    BALLOT ||--o{ VOTE : contains
    BALLOT }o--|| DOCUMENT : relates_to
    BALLOT }o--|| COMMITTEE : from
    
    VOTE }o--|| POLITICIAN : cast_by
    VOTE }o--|| PARTY : party_vote
    VOTE }o--|| BALLOT : in_ballot
    
    BALLOT {
        varchar ballot_id PK
        date ballot_date
        time ballot_time
        varchar issue
        varchar document_id FK
        varchar committee FK
        integer total_votes
        integer yes_count
        integer no_count
        integer abstain_count
        integer absent_count
    }
    
    VOTE {
        varchar vote_id PK
        varchar ballot_id FK
        varchar person_id FK
        varchar party FK
        varchar vote
        boolean is_rebel_vote
        boolean is_winning_vote
    }
    
    POLITICIAN {
        varchar person_id PK
        varchar party FK
        integer total_votes
        decimal annual_rebel_rate
        decimal annual_absence_rate
    }
    
    PARTY {
        varchar party_id PK
        decimal avg_win_rate
        decimal avg_discipline_score
    }

erDiagram
    DOCUMENT ||--o{ DOCUMENT_AUTHOR : has
    DOCUMENT ||--o{ DOCUMENT_ATTACHMENT : includes
    DOCUMENT ||--o{ DOCUMENT_REFERENCE : references
    DOCUMENT ||--o{ COMMITTEE_DECISION : leads_to
    DOCUMENT }o--|| COMMITTEE : assigned_to
    
    DOCUMENT_AUTHOR }o--|| POLITICIAN : authored_by
    DOCUMENT_AUTHOR }o--|| DOCUMENT : for_document
    
    COMMITTEE_DECISION }o--|| DOCUMENT : about
    COMMITTEE_DECISION }o--|| COMMITTEE : made_by
    
    DOCUMENT {
        varchar document_id PK
        varchar document_type
        varchar rm
        date published_date
        varchar status
        varchar organ FK
        text title
        text subtitle
        text fulltext
    }
    
    DOCUMENT_AUTHOR {
        varchar document_id PK_FK
        varchar person_id PK_FK
        integer author_order
    }
    
    DOCUMENT_ATTACHMENT {
        varchar attachment_id PK
        varchar document_id FK
        varchar filename
        varchar url
        integer size_bytes
    }
    
    DOCUMENT_REFERENCE {
        varchar source_doc_id PK_FK
        varchar target_doc_id PK_FK
        varchar reference_type
    }
    
    COMMITTEE_DECISION {
        varchar decision_id PK
        varchar document_id FK
        varchar committee_id FK
        date decision_date
        varchar decision_outcome
        text decision_text
    }
    
    COMMITTEE {
        varchar committee_id PK
        varchar committee_name
        integer total_documents
    }
    
    POLITICIAN {
        varchar person_id PK
        integer total_documents
    }

erDiagram
    GOVERNMENT ||--o{ MINISTRY : contains
    GOVERNMENT }o--|| PARTY : led_by
    
    MINISTRY ||--o{ GOVERNMENT_ROLE : has_roles
    MINISTRY }o--|| POLITICIAN : headed_by
    MINISTRY }o--|| PARTY : party_affiliation
    
    GOVERNMENT_ROLE }o--|| POLITICIAN : held_by
    GOVERNMENT_ROLE }o--|| MINISTRY : part_of
    
    GOVERNMENT {
        varchar government_id PK
        varchar government_name
        date start_date
        date end_date
        varchar prime_minister_id FK
        varchar leading_party FK
        varchar coalition_parties
    }
    
    MINISTRY {
        varchar ministry_id PK
        varchar government_id FK
        varchar ministry_name
        varchar minister_person_id FK
        varchar party FK
        integer decision_count
        decimal effectiveness_score
        varchar risk_level
    }
    
    GOVERNMENT_ROLE {
        varchar role_id PK
        varchar role_name
        varchar role_type
        varchar ministry_id FK
        varchar person_id FK
        varchar party FK
        date start_date
        date end_date
    }
    
    POLITICIAN {
        varchar person_id PK
        varchar party FK
    }
    
    PARTY {
        varchar party_id PK
        varchar party_name
    }

erDiagram
    POLITICIAN ||--o{ RULE_VIOLATION : has
    POLITICIAN ||--|| RISK_ASSESSMENT : assessed_by
    
    PARTY ||--o{ POLITICIAN : contains
    PARTY ||--|| PARTY_RISK_METRICS : has_metrics
    
    MINISTRY ||--|| MINISTRY_RISK_METRICS : has_metrics
    
    RULE_VIOLATION {
        varchar violation_id PK
        varchar person_id FK
        date violation_date
        varchar violation_type
        text violation_description
        varchar severity
    }
    
    RISK_ASSESSMENT {
        varchar person_id PK_FK
        decimal risk_score
        varchar risk_level
        integer total_violations
        decimal annual_absence_rate
        decimal annual_rebel_rate
        date last_updated
        text risk_assessment_text
    }
    
    PARTY_RISK_METRICS {
        varchar party_id PK_FK
        integer members_at_risk
        decimal avg_party_risk_score
        integer total_party_violations
        varchar party_risk_level
    }
    
    MINISTRY_RISK_METRICS {
        varchar ministry_id PK_FK
        varchar risk_level
        integer risk_violations
        decimal risk_score
        date last_assessment
    }
    
    POLITICIAN {
        varchar person_id PK
        varchar party FK
        decimal risk_score
        varchar risk_level
    }
    
    PARTY {
        varchar party_id PK
    }
    
    MINISTRY {
        varchar ministry_id PK
    }

URL: https://data.riksdagen.se/
Type: REST API + Open Data Portal
Authentication: None (public data)
Data Coverage: 1971-present (50+ years)

Endpoints:

• /personlista/ - MPs and politicians
• /dokument/ - Parliamentary documents
• /votering/ - Voting records
• /utskott/ - Committee information
• /anforande/ - Chamber speeches

Update Frequency:

• Real-time: Votes, speeches
• Daily: Documents, committee reports
• On change: MP assignments, party membership

Data Completeness: 98.5% (estimated)
Reliability: 99.9% uptime (government infrastructure)

Integration Method: CIA Platform batch processing + real-time updates

URL: https://val.se/
Type: Open Data Portal
Authentication: None (public data)
Data Coverage: 1911-present (electoral results)

Data Products:

• Election results (Riksdag, kommun, region)
• Voter turnout statistics
• Electoral district boundaries
• Candidate lists

Update Frequency: Post-election (every 4 years + by-elections)
Data Format: CSV, Excel, JSON
Reliability: Official government source

Integration Method: Manual download + CIA Platform import

URL: https://www.esv.se/psidata/
Type: PSI Data Portal (Public Sector Information)
Authentication: None (public data)
Data Coverage: 2000-present (budget data)

Data Products:

• Government budget (Statsbudget)
• Ministry spending (Utgiftsområden)
• Agency budgets
• Financial forecasts

Update Frequency: Annual (budget cycle) + quarterly reports
Data Format: CSV, Excel
Reliability: Official government financial data

Integration Method: Manual download + CIA Platform import

URL: https://data.worldbank.org/
Type: REST API + Open Data Portal
Authentication: None (public data)
Data Coverage: 1960-present (country indicators)

Data Products:

• GDP per capita
• Government effectiveness indicators
• Education and health metrics
• Democracy indices

Update Frequency: Annual
Data Format: JSON, CSV, XML
Reliability: International organization standard

Integration Method: CIA Platform API client

URL: https://data.imf.org/ (documentation) • www.imf.org/external/datamapper/api/v1 (Datamapper JSON) • api.imf.org/external/sdmx/3.0 (SDMX 3.0)
Type: REST (Datamapper JSON v1) + SDMX 3.0
Authentication: None (public data)
Data Classification: Public (same as SCB / World Bank)
Data Coverage: 1980-present (macro); annual + quarterly + monthly depending on dataset; projections to ~2031

Data Products:

• WEO (World Economic Outlook) — NGDP_RPCH (real GDP growth), PCPIPCH (CPI inflation), LUR (unemployment), GGXWDG_NGDP (gross debt / GDP), BCA_NGDPD (current account / GDP), …
• Fiscal Monitor (FM) — fiscal balance, primary balance, expenditure composition
• IFS (International Financial Statistics) — monetary, FX, balance-of-payments series
• MFS (Monetary & Financial Statistics) — policy rate, money-market rates
• GFS_COFOG — committee-aligned government spending by function
• DOTS (Direction of Trade Statistics) — bilateral trade flows
• ~155 SDMX databases in total

Update Frequency:

• WEO: April and October each year (projections refreshed at each vintage)
• Fiscal Monitor: April and October
• IFS: monthly
• MFS: monthly
• Projections published at T+5 years per vintage

Data Format: JSON (Datamapper), SDMX 3.0 JSON / XML
Reliability: ~99.5% availability; international-organization standard

Integration Method: Pure-TypeScript client scripts/imf-client.ts (sibling of scripts/world-bank-client.ts and scripts/scb-client.ts) — not an MCP server (ADR 0001 rationale: npm-SBOM coverage, no Python / uvx / third-party MCP). Invoked by agentic workflows via the bash tool and imported directly by build-time scripts.

Schema/Validation: DatamapperResponse shape in imf-client.ts (numeric-finite check, year parse-guard); SDMX 3.0 schema validation for structural metadata.

Caching: analysis/data/imf/{indicator}/{country}.json + sidecar .meta.json (mcpTool: imf-ts-client, projectionVintage: "WEO-2026-04", fetch timestamp).

Rate-limit handling: ~10 req / 5 s, 3× exponential back-off (1s → 2s → 4s), multi-country batching via Datamapper compare.

Allowlisted egress hosts: data.imf.org, api.imf.org, www.imf.org.

Supporting docs (referenced, not duplicated): analysis/imf/README.md, analysis/imf/indicator-policy-mapping.md, analysis/imf/use-cases.md, docs/adr/0001-adopt-imf-data-alongside-world-bank.md, .github/aw/ECONOMIC_DATA_CONTRACT.md.

URL: https://www.hack23.com/cia
Type: Java/Spring Boot application (backend data processing)
Authentication: Public read access, admin for updates
Data Coverage: Aggregated data from all sources above

Purpose:

• Data aggregation and normalization
• Intelligence product generation
• Risk assessment calculations
• Historical trend analysis

Update Frequency: Daily batch processing (03:00 CET)
Output Format: CSV exports, JSON statistics
Reliability: 99% uptime (self-hosted)

Integration Method: Direct CSV export consumption by Riksdagsmonitor

Quality Assurance:

• Automated validation against JSON schemas
• Cross-reference checks between sources
• Anomaly detection on data imports
• Manual spot-checking of critical data
• Version control of all data files

Location: /schemas/cia/
Purpose: Validate CIA platform data exports
Standard: JSON Schema Draft 2020-12

Schema Example (party-performance.schema.json):

{
  "$schema": "https://json-schema.org/draft/2020-12/schema",
  "$id": "https://riksdagsmonitor.com/schemas/cia/party-performance.schema.json",
  "title": "Party Performance Schema",
  "description": "Validates party effectiveness and performance metrics",
  "type": "object",
  "properties": {
    "party": {
      "type": "string",
      "enum": ["S", "M", "SD", "C", "V", "MP", "KD", "L"]
    },
    "year": {
      "type": "integer",
      "minimum": 1971,
      "maximum": 2026
    },
    "performance_score": {
      "type": "number",
      "minimum": 0,
      "maximum": 100
    },
    "win_rate": {
      "type": "number",
      "minimum": 0,
      "maximum": 100
    }
  },
  "required": ["party", "year", "performance_score"]
}

Location: /cia-data/
Format: UTF-8 encoded CSV with header row
Delimiter: Comma (,)
Quote Character: Double quote (")
Line Ending: LF (\n)

Header Row Requirements:

• First row must contain field names
• Field names: lowercase with underscores (snake_case)
• No spaces or special characters (except underscore)

Data Type Conventions:

• Dates: YYYY-MM-DD (ISO 8601)
• Timestamps: YYYY-MM-DD HH:MM:SS (ISO 8601)
• Decimals: Dot separator (.), 2 decimal places
• Booleans: true, false (lowercase)
• NULL values: Empty field (no quotes)

Example (politician risk summary):

person_id,first_name,last_name,party,status,risk_score,risk_level
0665485817222,Daniel,Bäckström,C,Tjänstgörande riksdagsledamot,50.00,HIGH
0836001490919,Ulrika,Heie,C,Tjänstgörande riksdagsledamot,38.00,MEDIUM

File: /cia-data/production-stats.json
Purpose: Daily statistics from CIA platform
Update: Daily at 03:00 CET via GitHub Actions

Schema Structure:

{
  "metadata": {
    "source_url": "string (URL)",
    "last_updated": "string (ISO 8601 timestamp)",
    "extraction_time": "string (ISO 8601 timestamp)",
    "generated_at": "string (ISO 8601 timestamp)",
    "version": "string (semantic version)"
  },
  "counts": {
    "total_persons": "integer",
    "total_votes": "integer",
    "total_documents": "integer",
    "total_committee_documents": "integer",
    "total_rule_violations": "integer",
    "total_political_parties": "integer",
    "total_governments": "integer",
    "total_government_roles": "integer",
    "total_government_role_members": "integer"
  },
  "tables": {
    "success": [
      {
        "name": "string (table name)",
        "count": "integer (row count)"
      }
    ],
    "empty": ["string (table name)"]
  }
}

Validation Rules:

• metadata.last_updated must be within 48 hours
• counts.* must be non-negative integers
• tables.success[].count must be positive
• metadata.version must follow semantic versioning

File: /cia-data/data-manifest.json
Purpose: Track all CSV files, checksums, and field descriptions
Update: On data file changes

Schema Structure:

{
  "version": "1.0.0",
  "last_updated": "2026-02-15",
  "files": [
    {
      "path": "politician/view_politician_risk_summary_sample.csv",
      "size_bytes": 69755,
      "checksum_sha256": "abc123...",
      "record_count": 349,
      "fields": [
        {
          "name": "person_id",
          "type": "string",
          "description": "Unique person identifier",
          "required": true,
          "example": "0665485817222"
        }
      ]
    }
  ]
}

Workflow: .github/workflows/validate-data.yml (if implemented)
Trigger: On push to cia-data/ directory
Steps:

1. Validate JSON files against schemas
2. Check CSV headers and data types
3. Verify production-stats.json freshness
4. Run integrity checks (foreign key validation)
5. Generate validation report

Exit Codes:

• 0: All validations passed
• 1: Schema validation failed
• 2: Data integrity check failed
• 3: Freshness check failed

Location: Dashboard JavaScript files
Method: Papa Parse CSV validation
Timing: On data load

Validation Checks:

• Header row presence
• Required field presence
• Data type validation (dates, numbers)
• Range validation (scores 0-100)
• Enum validation (party codes, risk levels)

Error Handling:

• Invalid data: Skip row, log warning
• Missing file: Fallback to remote URL
• Parse error: Display user-friendly message

Versioning Scheme: Semantic Versioning (MAJOR.MINOR.PATCH)

• MAJOR: Breaking changes (field removal, type change)
• MINOR: Additive changes (new fields, optional)
• PATCH: Documentation updates, bug fixes

Compatibility:

• Dashboards support current MAJOR version
• Backward compatibility for 1 MINOR version
• Deprecation notice: 6 months before MAJOR change

Schema Evolution Process:

1. Propose schema change (GitHub issue)
2. Update schema file with new version
3. Test with sample data
4. Update dashboard code (if breaking change)
5. Deploy schema, then data
6. Deprecation notice for old schema (if applicable)

graph TB
    subgraph "External Sources"
        Riksdag[Riksdag API<br/>data.riksdagen.se]
        Election[Election Authority<br/>val.se]
        Finance[Financial Authority<br/>esv.se]
        WorldBank[World Bank<br/>data.worldbank.org]
        IMF[IMF<br/>data.imf.org / api.imf.org<br/>WEO + SDMX 3.0]
    end
    
    subgraph "CIA Platform (Backend)"
        ETL[ETL Processes<br/>Spring Batch Jobs]
        DB[(PostgreSQL<br/>Production Database)]
        Analytics[Analytics Engine<br/>Risk Assessment]
        Export[CSV Export<br/>Sample Data]
    end
    
    subgraph "Riksdagsmonitor (Frontend)"
        GitHub[GitHub Repository<br/>cia-data/]
        LocalCache[Browser LocalStorage<br/>1-24 hour TTL]
        Dashboard[Interactive Dashboards<br/>Chart.js + D3.js]
    end
    
    subgraph "Content Delivery"
        CloudFront[AWS CloudFront<br/>Primary CDN]
        S3[S3 Storage<br/>Multi-region]
        GitHubPages[GitHub Pages<br/>DR Hosting]
    end
    
    Riksdag -->|REST API| ETL
    Election -->|CSV Download| ETL
    Finance -->|CSV Download| ETL
    WorldBank -->|REST API| ETL
    IMF -->|Datamapper JSON + SDMX 3.0<br/>pure-TS client, no MCP| ETL
    
    ETL --> DB
    DB --> Analytics
    Analytics --> DB
    DB --> Export
    
    Export -->|Daily 03:00 CET| GitHub
    GitHub -->|Deploy| CloudFront
    GitHub -->|Deploy| GitHubPages
    
    CloudFront -->|HTTPS/TLS 1.3| Dashboard
    GitHubPages -.->|DR Failover| Dashboard
    
    Dashboard -->|Load Data| LocalCache
    LocalCache -->|Cache Hit| Dashboard
    LocalCache -->|Cache Miss| GitHub
    
    style Riksdag fill:#ff9800,color:#000000
    style ETL fill:#4caf50,color:#000000
    style DB fill:#2196f3,color:#ffffff
    style Export fill:#9c27b0,color:#ffffff
    style GitHub fill:#ff9800,color:#000000
    style CloudFront fill:#4caf50,color:#000000
    style Dashboard fill:#00bcd4,color:#000000

Workflow: .github/workflows/update-cia-stats.yml
Schedule: Daily at 03:00 CET (02:00 UTC)
Trigger: cron: '0 2 * * *'

Steps:

1. Fetch Production Stats

   curl https://raw.githubusercontent.com/Hack23/cia/master/service.data.impl/sample-data/extraction_summary_report.csv

2. Parse CSV (Papa Parse)
3. Generate JSON (cia-data/production-stats.json)
4. Update Website Files (inject stats into HTML)
5. Git Commit (automated commit with stats)
6. Deploy (push to main → triggers deployment)

Error Handling:

• Network failure: Retry 3 times with exponential backoff
• Parse error: Log error, skip update, alert maintainer
• Stale data: Accept if <48 hours old, alert if older

Workflow: Weekly schema validation check
Trigger: cron: '0 0 * * 0' (Sundays at midnight)

Steps:

1. Fetch CIA Schemas (from CIA repo)
2. Compare with Local Schemas (schemas/cia/)
3. Detect Changes (field additions, type changes)
4. Generate Report (Markdown diff)
5. Create Issue (if changes detected)

Change Types:

• Additive: New optional fields (auto-accept)
• Deprecation: Fields marked deprecated (6-month notice)
• Breaking: Field removal or type change (manual review)

Location: Browser LocalStorage
Scope: Per-origin (https://riksdagsmonitor.com)
Capacity: ~10MB per origin (browser-dependent)

Cache Key Pattern:

riksdagsmonitor_cache_{data_type}_{language}

Example:

localStorage.setItem('riksdagsmonitor_cache_politician_risk_en', JSON.stringify({
  timestamp: Date.now(),
  ttl: 3600000, // 1 hour in milliseconds
  data: csvData
}));

TTL (Time-To-Live):

• Real-time data (votes): 5 minutes
• Daily data (documents): 1 hour
• Weekly data (risk assessments): 24 hours
• Monthly data (historical trends): 7 days

Cache Invalidation:

• Expiration: Automatic when TTL exceeded
• Manual: User refresh action (Ctrl+R)
• Version: On schema version change

CDN: GitHub Pages built-in CDN
Cache-Control Headers: Set by GitHub Pages
Default TTL: 10 minutes

Cacheable Assets:

• HTML pages: 10 minutes
• CSS files: 1 hour
• JavaScript files: 1 hour
• CSV data files: 10 minutes
• Images: 1 day

Cache Busting:

• Method: Git commit SHA in deployment
• Pattern: Files served from main branch HEAD
• Invalidation: Automatic on new deployment

CDN: AWS CloudFront (primary)
Edge Locations: 600+ globally
Default TTL: 86400 seconds (24 hours)

Cache Behaviors:

Path Pattern         TTL      Cache-Control
/                    3600s    public, max-age=3600
*.html               3600s    public, max-age=3600
*.css                86400s   public, max-age=86400
*.js                 86400s   public, max-age=86400
/cia-data/*.csv      3600s    public, max-age=3600
/cia-data/*.json     3600s    public, max-age=3600

Invalidation:

• Manual: AWS CLI invalidation command
• Automatic: On deployment (via GitHub Actions)
• Pattern: /* (all files)
• Cost: First 1,000 invalidations/month free

Implementation: JavaScript function in dashboards

Freshness Criteria:

function isDataFresh(timestamp, ttl) {
  const now = Date.now();
  const age = now - timestamp;
  return age < ttl;
}

const FRESHNESS_TTL = {
  realtime: 5 * 60 * 1000,      // 5 minutes
  daily: 60 * 60 * 1000,         // 1 hour
  weekly: 24 * 60 * 60 * 1000,   // 24 hours
  monthly: 7 * 24 * 60 * 60 * 1000 // 7 days
};

Fallback Strategy:

1. Check LocalStorage (cache)
2. If fresh: Use cached data
3. If stale: Fetch from GitHub (cia-data/)
4. If GitHub fails: Fallback to remote URL
5. If all fail: Display error message

graph LR
    A[Collection<br/>CIA Platform] --> B[Validation<br/>Schema Checks]
    B --> C[Storage<br/>GitHub + S3]
    C --> D[Presentation<br/>Dashboards]
    
    B -->|Invalid Data| E[Error Log]
    E -->|Alert| F[Maintainer]
    
    C -->|Cache| G[LocalStorage]
    G -->|Serve| D
    
    style A fill:#ff9800,color:#000000
    style B fill:#4caf50,color:#000000
    style C fill:#2196f3,color:#ffffff
    style D fill:#9c27b0,color:#ffffff
    style E fill:#f44336,color:#ffffff

Pipeline Stages:

1. Collection (CIA Platform)

   • Source: Riksdag API, Election Authority, etc.
   • Frequency: Real-time to daily
   • Output: PostgreSQL database

2. Validation (CIA Platform + GitHub Actions)

   • Schema validation (JSON Schema)
   • Data integrity checks (foreign keys)
   • Completeness checks (required fields)
   • Output: Validated CSV files

3. Storage (GitHub + AWS)

   • Primary: GitHub repository (cia-data/)
   • Secondary: AWS S3 multi-region
   • Backup: Git history (immutable)

4. Presentation (Riksdagsmonitor)

   • Load: LocalStorage cache → GitHub → Remote
   • Parse: Papa Parse CSV parser
   • Render: Chart.js + D3.js visualizations

Riksdagsmonitor supports 14 languages with full data model localization:

Politicians (person_data):

• Names: Original Swedish (no translation)
• Party: Translated party abbreviation context
• Status: Translated status values

Parties (sweden_political_party):

{
  "party_id": "S",
  "translations": {
    "en": "Social Democrats",
    "sv": "Socialdemokraterna",
    "de": "Sozialdemokraten",
    "fr": "Sociaux-démocrates",
    "es": "Socialdemócratas",
    "ar": "الديمقراطيون الاجتماعيون",
    "ja": "社会民主党",
    "zh": "社会民主党"
  }
}

Committees (committee_document_data):

{
  "committee_id": "AU",
  "translations": {
    "en": "Labour Market Committee",
    "sv": "Arbetsmarknadsutskottet",
    "de": "Ausschuss für Arbeitsmarkt",
    "fr": "Commission du marché du travail",
    "es": "Comisión del Mercado Laboral",
    "ar": "لجنة سوق العمل",
    "ja": "労働市場委員会",
    "zh": "劳动市场委员会"
  }
}

Document Types:

{
  "document_type": "mot",
  "translations": {
    "en": "Motion (MP-initiated proposal)",
    "sv": "Motion",
    "de": "Antrag",
    "fr": "Motion",
    "es": "Moción",
    "ar": "اقتراح",
    "ja": "動議",
    "zh": "动议"
  }
}

Dashboard Titles:

• Stored in HTML <title> tags per language
• Pattern: index_{lang}.html

Chart Labels:

• Translated in JavaScript dashboard code
• Language detection: document.documentElement.lang

Data Classification Values:

{
  "risk_level": {
    "LOW": {
      "en": "Low Risk",
      "sv": "Låg risk",
      "de": "Geringes Risiko",
      "ar": "مخاطر منخفضة"
    },
    "MEDIUM": {
      "en": "Medium Risk",
      "sv": "Medelhög risk",
      "de": "Mittleres Risiko",
      "ar": "مخاطر متوسطة"
    },
    "HIGH": {
      "en": "High Risk",
      "sv": "Hög risk",
      "de": "Hohes Risiko",
      "ar": "مخاطر عالية"
    }
  }
}

Affected Languages: Arabic (ar), Hebrew (he)

HTML Structure:

<html lang="ar" dir="rtl">
  <head>
    <meta charset="UTF-8">
    <title>Riksdagsmonitor - مراقب البرلمان السويدي</title>
  </head>
  <body>
    <!-- Content flows right-to-left -->
  </body>
</html>

CSS Adaptations:

[dir="rtl"] .dashboard {
  text-align: right;
  direction: rtl;
}

[dir="rtl"] .chart-legend {
  float: left; /* Reversed from LTR */
}

Data Implications:

• Text fields: Unicode support required
• Sorting: Locale-aware sorting
• Rendering: Browser handles RTL layout

Pattern: index_{lang}.html

Files (14 languages):

index.html       (English - default)
index_sv.html    (Swedish)
index_da.html    (Danish)
index_no.html    (Norwegian)
index_fi.html    (Finnish)
index_de.html    (German)
index_fr.html    (French)
index_es.html    (Spanish)
index_nl.html    (Dutch)
index_ar.html    (Arabic - RTL)
index_he.html    (Hebrew - RTL)
index_ja.html    (Japanese)
index_ko.html    (Korean)
index_zh.html    (Chinese)

Sitemap Files:

sitemap.xml      (Master sitemap)
sitemap.html     (English - default)
sitemap_sv.html  (Swedish)
sitemap_ar.html  (Arabic)
...

Purpose: Signal alternate language versions to search engines

Implementation:

<head>
  <link rel="alternate" hreflang="en" href="https://riksdagsmonitor.com/index.html" />
  <link rel="alternate" hreflang="sv" href="https://riksdagsmonitor.com/index_sv.html" />
  <link rel="alternate" hreflang="da" href="https://riksdagsmonitor.com/index_da.html" />
  <link rel="alternate" hreflang="no" href="https://riksdagsmonitor.com/index_no.html" />
  <link rel="alternate" hreflang="fi" href="https://riksdagsmonitor.com/index_fi.html" />
  <link rel="alternate" hreflang="de" href="https://riksdagsmonitor.com/index_de.html" />
  <link rel="alternate" hreflang="fr" href="https://riksdagsmonitor.com/index_fr.html" />
  <link rel="alternate" hreflang="es" href="https://riksdagsmonitor.com/index_es.html" />
  <link rel="alternate" hreflang="nl" href="https://riksdagsmonitor.com/index_nl.html" />
  <link rel="alternate" hreflang="ar" href="https://riksdagsmonitor.com/index_ar.html" />
  <link rel="alternate" hreflang="he" href="https://riksdagsmonitor.com/index_he.html" />
  <link rel="alternate" hreflang="ja" href="https://riksdagsmonitor.com/index_ja.html" />
  <link rel="alternate" hreflang="ko" href="https://riksdagsmonitor.com/index_ko.html" />
  <link rel="alternate" hreflang="zh" href="https://riksdagsmonitor.com/index_zh.html" />
  <link rel="alternate" hreflang="x-default" href="https://riksdagsmonitor.com/index.html" />
</head>

Benefits:

• Improved SEO for international audiences
• Correct language version served by search engines
• Reduced duplicate content penalties

Method: Manual language selection (no auto-detect)

Navigation:

• Language picker in website header
• Flags or language names as buttons
• Preserves dashboard state on language change

URL Pattern:

https://riksdagsmonitor.com/             → English (default)
https://riksdagsmonitor.com/index_sv.html → Swedish
https://riksdagsmonitor.com/index_ar.html → Arabic

Data Loading:

• Same CSV data files for all languages
• Translation layer in JavaScript dashboard code
• Locale-specific number/date formatting

Strategy: Client-side caching with TTL-based expiration

Implementation:

class DataCache {
  constructor(ttl = 3600000) { // 1 hour default
    this.ttl = ttl;
  }
  
  set(key, data) {
    const item = {
      data: data,
      timestamp: Date.now(),
      ttl: this.ttl
    };
    localStorage.setItem(key, JSON.stringify(item));
  }
  
  get(key) {
    const item = JSON.parse(localStorage.getItem(key));
    if (!item) return null;
    
    const age = Date.now() - item.timestamp;
    if (age > item.ttl) {
      localStorage.removeItem(key);
      return null;
    }
    
    return item.data;
  }
}

Cache Keys:

riksdagsmonitor_cache_politician_risk_en
riksdagsmonitor_cache_party_performance_sv
riksdagsmonitor_cache_seasonal_patterns_de
riksdagsmonitor_cache_election_cycle_fr

TTL Configuration:

• Real-time: 5 minutes (votes, live updates)
• Daily: 1 hour (documents, statistics)
• Weekly: 24 hours (risk assessments, trends)
• Historical: 7 days (longitudinal analysis)

GitHub Pages CDN: Built-in global CDN
Cache-Control Headers: Managed by GitHub

Effective Caching:

Cache-Control: max-age=600  (10 minutes)

Benefits:

• Global edge caching
• Reduced origin requests
• Faster load times (50-200ms typical)

Limitations:

• No custom cache headers
• No manual invalidation
• 10-minute minimum TTL

Freshness Criteria:

const FRESHNESS_POLICY = {
  production_stats: {
    ttl: 24 * 60 * 60 * 1000, // 24 hours
    acceptable_age: 48 * 60 * 60 * 1000 // 48 hours max
  },
  politician_risk: {
    ttl: 60 * 60 * 1000, // 1 hour
    acceptable_age: 24 * 60 * 60 * 1000 // 24 hours max
  },
  seasonal_patterns: {
    ttl: 7 * 24 * 60 * 60 * 1000, // 7 days
    acceptable_age: 30 * 24 * 60 * 60 * 1000 // 30 days max
  }
};

Stale Data Handling:

• Display age indicator: "Data updated 2 hours ago"
• Warning for old data: "Data may be outdated (>24 hours old)"
• Error for very old data: "Data stale (>48 hours), refresh needed"

Implementation: Load data on-demand per dashboard

Strategy:

// Load only when dashboard becomes visible
const observer = new IntersectionObserver((entries) => {
  entries.forEach(entry => {
    if (entry.isIntersecting) {
      loadDashboardData(entry.target.dataset.dashboard);
    }
  });
});

document.querySelectorAll('.dashboard').forEach(el => {
  observer.observe(el);
});

Benefits:

• Reduced initial page load time
• Lower bandwidth consumption
• Better user experience (faster FCP)

Build System: Vite 7 with ES modules

Split Points:

js/election-cycle-dashboard.js   (46KB)
js/party-dashboard.js            (TBD)
js/seasonal-patterns-dashboard.js (TBD)
scripts/committees-dashboard.js   (39KB)
scripts/coalition-dashboard.js    (33KB)

Loading Strategy:

// Dynamic import per dashboard
if (document.getElementById('election-cycle-dashboard')) {
  import('./js/election-cycle-dashboard.js')
    .then(module => module.init())
    .catch(err => console.error('Failed to load dashboard', err));
}

Benefits:

• Smaller initial bundle size
• Faster time to interactive (TTI)
• Better Core Web Vitals scores

Measurement Tools:

• Lighthouse CI (automated)
• WebPageTest (manual)
• Chrome DevTools Performance panel

File: /budget.json

Resource Size Budgets:

{
  "document": "105KB",     // HTML pages
  "stylesheet": "370KB",   // CSS files
  "script": "300KB",       // JavaScript bundles
  "image": "500KB",        // Images
  "font": "100KB",         // Web fonts
  "total": "1200KB"        // Total page weight
}

Performance Budgets:

{
  "first-contentful-paint": "5100ms",
  "largest-contentful-paint": "5400ms",
  "interactive": "5400ms",
  "cumulative-layout-shift": "0.1",
  "total-blocking-time": "200ms"
}

Enforcement: GitHub Actions workflow fails on budget violations

graph TB
    subgraph "External Data Sources"
        Riksdag[Swedish Riksdag API<br/>data.riksdagen.se<br/>2.5M votes, 109K docs]
        Election[Election Authority<br/>val.se<br/>Electoral results]
        Finance[Financial Authority<br/>esv.se<br/>Budget data]
        WorldBank[World Bank<br/>data.worldbank.org<br/>Country indicators]
        IMF[IMF<br/>data.imf.org / api.imf.org<br/>WEO + Fiscal Monitor + IFS<br/>macro/fiscal + T+5 projections]
    end
    
    subgraph "Riksdagsmonitor System"
        System[Data Model<br/>2,494 politicians<br/>40 parties<br/>15 committees]
    end
    
    subgraph "CIA Platform (External)"
        CIA[CIA Backend<br/>Data aggregation<br/>Risk assessment]
    end
    
    subgraph "Users"
        Analyst[Political Analysts]
        Journalist[Journalists]
        Citizen[Citizens]
        Researcher[Researchers]
    end
    
    Riksdag -->|REST API| CIA
    Election -->|CSV Data| CIA
    Finance -->|CSV Data| CIA
    WorldBank -->|REST API| CIA
    IMF -->|Pure-TS client<br/>Datamapper JSON + SDMX 3.0<br/>no MCP| CIA
    
    CIA -->|CSV Exports| System
    CIA -->|Daily Statistics| System
    
    System -->|Visualizations| Analyst
    System -->|Visualizations| Journalist
    System -->|Visualizations| Citizen
    System -->|Visualizations| Researcher
    
    style Riksdag fill:#ff9800,color:#000000
    style Election fill:#ff9800,color:#000000
    style Finance fill:#ff9800,color:#000000
    style WorldBank fill:#ff9800,color:#000000
    style IMF fill:#00897b,color:#ffffff
    style CIA fill:#9c27b0,color:#ffffff
    style System fill:#4caf50,color:#000000
    style Analyst fill:#2196f3,color:#ffffff

graph TB
    subgraph "Data Storage Layer"
        GitHub[GitHub Repository<br/>cia-data/<br/>50+ CSV files]
        S3[AWS S3<br/>Multi-region<br/>Primary storage]
        LocalStorage[Browser LocalStorage<br/>Client-side cache<br/>~10MB capacity]
    end
    
    subgraph "Data Processing Layer"
        CIA_DB[(CIA Platform<br/>PostgreSQL Database<br/>Production data)]
        Export[CSV Export Engine<br/>Sample data generator]
        Stats[Statistics Generator<br/>production-stats.json]
    end
    
    subgraph "Validation Layer"
        Schemas[JSON Schemas<br/>2 schema files]
        Validator[Schema Validator<br/>GitHub Actions]
    end
    
    subgraph "Presentation Layer"
        Dashboards[Interactive Dashboards<br/>Chart.js + D3.js<br/>5 active dashboards]
    end
    
    CIA_DB --> Export
    CIA_DB --> Stats
    Export --> GitHub
    Stats --> GitHub
    
    GitHub --> Validator
    Schemas --> Validator
    Validator -->|Valid| S3
    
    S3 --> Dashboards
    GitHub --> Dashboards
    Dashboards --> LocalStorage
    LocalStorage --> Dashboards
    
    style CIA_DB fill:#2196f3,color:#ffffff
    style GitHub fill:#ff9800,color:#000000
    style S3 fill:#4caf50,color:#000000
    style Dashboards fill:#9c27b0,color:#ffffff
    style Validator fill:#f44336,color:#ffffff

This DATA_MODEL.md complements ARCHITECTURE.md:

ARCHITECTURE.md Focus:

• System architecture (infrastructure, deployment)
• Component interactions (CDN, hosting, CI/CD)
• Security architecture (HTTPS, CSP, access control)

DATA_MODEL.md Focus (this document):

• Data entities and relationships
• Data schemas and validation
• Data pipelines and caching
• Multi-language data structure

Cross-References:

• ARCHITECTURE.md Section 5 → DATA_MODEL.md Section 4 (Data Sources)
• ARCHITECTURE.md Section 3.1 → DATA_MODEL.md Section 3 (ERD Integration)
• ARCHITECTURE.md Section 8 → DATA_MODEL.md Section 8 (Performance)

Annex A.8 - Asset Management

Compliance Level: ✅ FULLY COMPLIANT

PR.DS - Data Security

Compliance Level: ✅ FULLY COMPLIANT

Control 3 - Data Protection

Compliance Level: ✅ FULLY COMPLIANT

Personal Data Processing:

Data Categories:

• Public Officials: MP names, roles, voting records
• Legal Basis: Article 6(1)(e) - Public interest
• Scope: Public figures acting in official capacity only

Data Subject Rights:

• Right to Access: All data publicly accessible
• Right to Rectification: Source data from official government APIs
• Right to Erasure: Not applicable (public interest exception)
• Right to Object: Not applicable (public interest exception)

Privacy-by-Design:

• No special category data (Article 9)
• No data about private individuals
• No user tracking or analytics
• No cookies or personal identifiers

Compliance Level: ✅ FULLY COMPLIANT (Public Interest Processing)

• ARCHITECTURE.md - System architecture and infrastructure
• README.md - Project overview and features
• SECURITY_ARCHITECTURE.md - Security controls
• THREAT_MODEL.md - STRIDE threat analysis
• TRANSLATION_GUIDE.md - Multi-language standards

• Hack23 ISMS - Information Security Management System
• Hack23 Public ISMS - Public security policies
• Secure Development Policy
• Classification Framework

• 🏛️ Architecture - C4 system architecture models
• 🔄 Flowchart - Data pipelines and workflow processes
• 📈 State Diagram - System state transitions
• 💼 SWOT - Strategic analysis and positioning
• 🎯 Threat Model - Security threat analysis
• 🔐 Security Architecture - Current security controls
• 🚀 Future Data Model - Future data architecture roadmap
• 🛠️ Secure Development Policy - Hack23 SDLC requirements
• 🏷️ Classification Framework - Data classification standards

• CIA Platform - Citizen Intelligence Agency backend
• CIA Data Specs - Export specifications
• Swedish Riksdag API - Official parliament API
• JSON Schema - Schema validation standard

📋 Document Control:
✅ Approved by: James Pether Sörling, CEO
📤 Distribution: Public
🏷️ Classification:
📅 Effective Date: 2026-02-15
⏰ Next Review: 2027-02-15
🎯 Framework Compliance:

Statskontoret adds a public Swedish-administration data domain under the economic/public-administration context layer.

analysis/data/statskontoret/{dataset}/{artifact}.json
analysis/data/statskontoret/{dataset}/{artifact}.meta.json

Sidecar metadata includes fetchedAt, mcpTool: statskontoret-ts-client, dataset, and artifact. The provider decision matrix in analysis/statskontoret/indicators-inventory.json maps government-body headcount and central-government budget outturn claims to Statskontoret, while macro/fiscal projections remain IMF-first.