M4 Master BundesEcho Bringing visibility to the political echo.

Automated Collection
BundesEcho continuously synchronizes with the source platform to ingest newly published citizen Q&A threads, keeping analyses up to date.

NLP Enrichment
Each entry is automatically analyzed and enriched with structured signals:

• Context: Education and professional background
• Topics: Assignment to core policy areas
• Rhetorical Clarity: Direct, Vague, or Evasive
• Positioning: Directional stance detection

Semantic Representation
All entries are embedded as semantic vectors, enabling meaning-based search and powering the RAG features.

Optimized Access
Enriched data and vectors are served via a high-performance API for fast, reliable access.

Directional Positioning
Political statements often go beyond simple pro or contra. BundesEcho applies a three-way stance model:

• Expanding: Advocates policy growth or support
• Limiting: Supports restriction or tighter regulation
• Status Quo: Defends existing policies

Explainable Metrics
Every AI-generated classification includes a short reasoning summary, ensuring transparency and interpretability.

Dashboard Visuals
The frontend uses Apache ECharts for interactive exploration:

• Sankey Diagrams: Visualize how questions flow from topics to parties
• Parliament View: Interactive Bundestag layout with sentiment and response-time insights
• Topic Sunburst: Drill-down from main categories to trending sub-topics
• Network Graph: Shows relationships between politicians, parties, and issues

Politician Profiles
Each member of the Bundestag has a dedicated analytics dashboard to assess their individual performance:

• Activity Timeline: A bar chart tracking the frequency of questions answered over time.
• Thematic Focus: A Pie Chart highlighting the “Top 5 Topics” addressed by the politician.
• Sentiment Gauge: A distinct visual meter showing the average rhetorical clarity and tonal sentiment of the politician’s answers.

Category Insights
Aggregated metrics per policy area include:

• Total questions, response rates, and average response times
• Party-level engagement comparisons

Semantic Search & AI Summaries
Users can search the dataset using natural language. Results are matched by meaning rather than keywords and summarized using AI, grounded in the original data.