Government Web Relationship Data

This page provides comprehensive insights into the web relationships between European and allied government websites. We analyze links, scripts, styles, forms, and media dependencies to understand how government sites connect and interact with each other and third-party services.

About This Dataset +

What is Extracted

This dataset contains relationship observations from government web pages scanned as part of the EU Government Accessibility Statement Discovery project. Relationships include:

• Editorial Links: <a href> links to internal/external pages
• Script Dependencies: <script src> external JavaScript resources
• Style Dependencies: <link rel="stylesheet"> CSS stylesheets
• Font/Prelooad Dependencies: <link rel="preload"|"font"|"preconnect"> font resources
• Image/Media Dependencies: <img src> images and media
• Form Destinations: <form action> external form submissions

Classification Categories

External domains are automatically categorized based on their function:

• Analytics: Google Analytics, Tag Manager, Heatmaps, etc.
• Social Platform: Facebook, Twitter/X, LinkedIn, Instagram, YouTube
• CDN: Content Delivery Networks like Cloudflare, Fastly, Akamai
• Identity: Authentication services like Okta, Auth0, Microsoft Login
• Document Host: File hosting services like Google Drive, Dropbox, SharePoint
• Commercial Service: Payment, support, and business services

Methodology Notes

• Graph Metrics: In-degree, out-degree, and weighted measures represent observed web relationships, NOT institutional authority or hierarchy.
• Country Assignment: Prioritized through institutional evidence, not just TLD mapping.
• Classification Precedence: Authoritative registry > Curated source > Institutional source > Structured source > Software Heritage > Domain pattern > Network signal
• Data Provenance: All data is sourced from machine-readable scan results with full tracking of observation timestamps and source pages.

Filters and Search

Country All Countries

Relationship Type All Types Editorial Link Script Dependency Style Dependency Font/Preload Image/Media Form Destination

Government Level All Levels Supranational National Regional State/Province Local Municipal Independent Public Body

Classification Status All Statuses Confirmed Probable Candidate Disputed

Target Category All Categories Analytics Social Platform CDN Identity Document Host Commercial Service Known Government Shared Government Service Unknown External

Search

Gov-to-Gov Links Only

Total Domains

0

Total Relationships

0

Countries

0

Unique Targets

0

Editorial Links

0

Technical Dependencies

0

Domains Table +

Country	Domain	Targets	Top Types	Observations	Last Seen
No domains match the current filters Try adjusting your search criteria or filters

Relationships Table +

Source Domain	Target Domain	Relationship Type	Source Pages	Observations	Page Regions	First Seen	Last Seen
No relationships match the current filters Try adjusting your search criteria or filters

Country Breakdown +

Loading country data...

Country Details

Most Connected Government Domains

Ranked by total observed connections for the selected country. Select a country above to see its ranking; select a domain to render its network in the graph below.

Government-to-Government Links

Editorial links where both source and target are government domains (excluding self-links). Use the "Gov-to-Gov Links Only" filter above to activate.

Filter by country: All Countries

Show full link table

Country: Select a country… Include third-party services All third-party types Scripts Stylesheets Fonts Images Forms

By default this shows government-to-government editorial links only for the selected country — no analytics, CDNs, or other third-party services. Pick a country above and select "Include third-party services" to see the rest.

Graph Notes: Node size = connection count. Edge thickness = observation count. This visualization represents observed web relationships and does NOT establish institutional authority or government hierarchy.

Downloads

Download the complete relationship dataset or specific filtered results for offline analysis.

📄 Download Complete JSON Dataset 📊 Download Complete CSV Dataset 📄 Download Filtered JSON 📊 Download Filtered CSV

Methodology and Limitations +

Data Extraction Process

All relationship data is extracted during the scanning process using the existing MultiScanner infrastructure. The HTML parsing is performed using BeautifulSoup and tldextract for domain normalization. URLs are resolved relative to their source page and normalized to eliminate fragments, default ports, and IDNA encoding.

Classification Logic

Domain categorization follows a configurable rule-based approach defined in data/relationship_categories.json. Categories are applied to registrable domains and represent preliminary indicators rather than definitive classifications.

Aggregation Methodology

For each scan batch, relationships are deduplicated within source pages and then aggregated by:

• Source registrable domain
• Target registrable domain
• Relationship type

Aggregation preserves evidence counts, timestamps, and page regions (main, navigation, footer, header, unknown).

Limitations and Interpretations

• Graph metrics (in-degree, out-degree) indicate observed connection patterns, not institutional authority or hierarchy.
• Domain-level analysis does not infer central government status from network structure.
• Country assignment prioritizes institutional evidence over TLD mapping alone.
• Classification uses preliminary categorization that may require manual verification for certain services.
• Data represents a snapshot in time; relationships may change as pages are updated.

Provenanced Data

All data maintains full provenance tracking with:

• Source page URLs
• Extraction timestamps
• Page region information
• Canonical domain records with classification basis
• Multiple source evidence for each domain

Submitting Corrections

If you identify errors in the relationship data, please use the issue tracker on GitHub to submit corrections. Include the source page URL, target domain, and the specific type of correction needed (classification, country assignment, relationship type, etc.).

Data Schema Version: 1.0.0
Last Updated: 2026-07-15

Domain Details