Overview

The InnWater Water Tariff Dashboard is an AI-augmented decision-support system designed to simulate, analyze, and optimize water pricing structures. Developed within the Horizon Europe InnWater project, it supports sustainable, equitable, and economically efficient tariff design in multi-level water governance contexts.

I contributed as AI Engineer/System Architect at Eurecat, integrating the AI Assistant layer and supporting backend architecture for economic simulation workflows.

💧 Water Tariff Simulation & Assessment

Water Tariff Simulation

Water Tariff Dashboard.

The dashboard enables structured tariff analysis through:

  • Simulation of alternative pricing schemes:
    • Flat rates
    • Increasing block tariffs
    • Progressive consumption models
  • Comparison between sanitation and non-sanitation subscribers
  • Comparison between poor and non-poor groups
  • Environmental and resource cost internalization

The tool converts tariff design into a reproducible, scenario-based analytical workflow.

🤖 AI Assistant – RAG-based Economic Interpretation

Innwater AI Assistant Diagram

Innwater AI Assistant Diagram

The Water Tariff Dashboard integrates a Retrieval-Augmented Generation (RAG) Agent architecture, shared across the InnWater platform.

Architecture

  • Query treatment module
  • Hierarchical retrieval over indexed project deliverables
  • Embedding-based semantic search
  • LLM-grounded response generation
  • Logging & evaluation via golden dataset benchmarking

Capabilities

  • Interprets tariff simulation outputs
  • Explains affordability indicators in policy terms
  • Highlights trade-offs between equity and cost recovery
  • Suggests scenario adjustments
  • Connects tariff outcomes with governance gaps and CGE model results
  • Multilingual interaction

The AI layer transforms the dashboard from a numerical simulator into an AI-assisted policy analysis tool.

🏗 System Architecture

Frontend

  • Angular 15
  • Bootstrap 5
  • Chart.js & D3.js
  • Interactive scenario comparison

Backend

  • FastAPI (Python)
  • PostgreSQL
  • SQLAlchemy ORM
  • Pandas & NumPy for economic modeling

AI Layer

  • RAG pipeline
  • Semantic embeddings
  • LLM response generation
  • Bias and robustness evaluation

Deployment

  • Docker-based containerization
  • Integrated within the InnWater Governance Platform

Impact

  • Operationalized water tariff theory into an AI-supported economic decision tool
  • Enabled evidence-based pricing design for water utilities
  • Bridged tariff modeling with governance and macroeconomic simulation
  • Demonstrated responsible AI deployment in public-sector economic policy