Skip to content

demo-script.md

Latest commit

 

History

History
120 lines (82 loc) · 6.83 KB

demo-script.md

File metadata and controls

120 lines (82 loc) · 6.83 KB
Raw

Demo Script

The following scenarios demonstrate how Climate Loop operates in realistic high risk situations. They simulate extreme weather events and community reports to validate the platform’s ability to turn complex climate data into clear and actionable guidance.

Each scenario shows how predictive models, structured emergency alerts, IoT environmental data, and AI generated explanations work together within a single operational flow.

At this stage, the functional prototype is focused on Lugo in Galicia, Spain. This defined geographic scope allows us to validate the full pipeline, from time series forecasting to alert processing and community contributions, within a realistic and controlled context.

⚠️ Important for Jury Testing: To ensure a consistent and accurate evaluation of the prototype, jurors are requested to keep the application’s location set to Lugo during testing. The current implementation is geographically scoped to Lugo, and changing the location may result in incomplete data visualization, inactive alerts, or map inconsistencies. All interactions, including viewing alerts, exploring the map, and submitting community reports, should be performed within the Lugo area.

These scenarios are intended to test the core premise of Climate Loop: that simplifying climate information can directly support faster and safer decision making in moments of risk.

Scenario 1 – Yellow Alert for High ultraviolet index forecast for Galicia

Context
An official Yellow Alert for intense and persistent rainfall is issued for Galicia and western Asturias.

Official Alert Input (CAP Structured Content) NIVEL NARANJA — PRECIPITACIONES INTENSAS Y PERSISTENTES

Emitido por: Agencia Estatal de Meteorología AEMET Fecha de emisión: 2/24/2026, 8:00 AM Área afectada: Comunidad Autónoma de Galicia y occidente del Principado de Asturias Vigencia: 2/26/2026, 11:00 AM – 2/26/2026, 05:00 PM

UV inusualmente alto para febrero en Galicia. Valores 7-8.

System Processing

  • The backend ingests and stores the structured CAP alert.
  • The system identifies it as the most recent active alert within its validity window.
  • Because it is currently active, it is automatically displayed on the home screen immediately after user login.
  • The AI layer generates a simplified explanation of the alert and leverages Generative AI to automatically produce and store multilingual translations in Spanish, English, French, and Portuguese.

AI Generated Explanation (Spanish Example) Análisis de la alerta Lluvia intensa prevista para toda Galicia y el oeste de Asturias con acumulados de hasta 80 mm en 12 horas. Las zonas costeras y montañosas pueden recibir cantidades superiores. Riesgo elevado de inundaciones repentinas en áreas urbanas y desbordamientos de ríos menores.

Toda la población de Galicia y occidente de Asturias, especialmente residentes en áreas urbanas propensas a inundaciones y cercanías de cauces fluviales.

Acciones recomendadas

1. Evitar desplazamientos por carretera, especialmente en zonas bajas y cercanas a ríos 2. Limpiar desagües y canaletas si es posible antes del inicio de la lluvia 3. Alejar vehículos de zonas susceptibles de inundación 4. Tener preparado un kit de emergencia con linterna, agua y documentos 5. Llamar al 112 en caso de emergencia

User Interaction Flow

  • Upon login, the user immediately sees the active Orange Alert highlighted on the home screen.
  • The explanation is automatically presented in the user’s selected language.
  • The user can share the alert and recommended actions.
  • The user can initiate a conversation with the AI assistant.
  • The assistant receives the active alert as contextual input and answers follow up questions about risk level, duration, affected areas, or recommended precautions.

Expected Outcome Users immediately understand that an active Orange Alert is in effect in their region.
They receive clear and localized guidance in their preferred language.
They can interact with the AI assistant for further clarification, strengthening situational awareness and informed decision making.

Scenario 2 – Interactive Risk Map with Official Alerts and Community Contributions

Context
The user wants to visualize active risks and local reports in their region.

User Action

  • The user logs into the application.
  • The user taps the bottom navigation bar and selects the “Map” icon.

System Behavior

  • The system loads an interactive map centered on the user’s region.
  • Official active alerts are displayed as categorized markers according to severity level.
  • Community contributions are displayed as separate markers with distinct visual identifiers.
  • A radius control allows the user to adjust the visible geographic range.
  • A legend explains alert levels, color codes, and contribution types.

When the user taps on a map marker:

  • If it is an official alert, the user is redirected to the full alert details and AI generated explanation.
  • If it is a community contribution, the user is redirected to the submitted report, including description and validated image.

Expected Outcome

Users gain a spatial understanding of ongoing risks in their area.
They can differentiate between official institutional alerts and community generated reports.
They can explore detailed information directly from the map interface, improving situational awareness and localized decision making.

Scenario 3 – Community Contribution Submission with AI Image Validation

Context
A user wants to report a climate related event occurring in their area.

User Action

  • The user taps the “Report” button in the application.
  • The user selects a category for the event, such as flooding, strong winds, or landslide risk.
  • The user provides a title for the report.
  • The user may optionally add descriptive notes with additional context.
  • The user selects or refines the event location using the address search API to specify the exact address.
  • The user captures a photo using the device camera.

System Behavior

  • The image is processed by the AI validation layer before any data is stored.
  • The system performs person detection to identify whether identifiable individuals are present in the image.
  • If a person is detected, the submission is automatically rejected.
  • In this case, neither the contribution nor the image is saved to the database.
  • If no person is detected and validation rules are satisfied, the contribution is stored and becomes visible on the map and in the community feed.

Expected Outcome

The platform enables community reporting while preserving privacy and preventing the storage of images containing identifiable individuals.
Only validated and compliant contributions are persisted, reinforcing a privacy by design and data minimization approach.