AFRAM

Posted on by Steven McDermott

Team Name

Fire Squad

Timeline

Fall 2024 – Spring 2025

Students

  • Laurencia Anzela – Computer Science
  • Biplav Aryal – Computer Science
  • Hien Dao – Computer Science
  • Johnny Gonzalez – Software Engineering
  • Jason Lu – Software Engineering
  • Khang Nguyen Computer Science

Sponsor

Arlington Fire Department

Abstract

The Arlington Fire Risk Assessment Map (AFRAM) is a predictive web platform that visualizes fire risk levels across the city of Arlington. The system integrates historical fire incident data, real-time weather inputs, and machine learning predictions to assist emergency preparedness. The map offers dynamic filtering, heatmaps, and visual indicators to help users assess risk patterns across neighborhoods.

Background

Arlington Fire Risk Assessment Map (AFRAM) is a machine learning-powered web application designed to help emergency responders and city planners proactively identify and mitigate fire hazards in the city of Arlington. By integrating real-time and historical data-including fire incident data, weather conditions, and building density-AFRAM generates predictive risk scores and visualizes them on an interactive map. These insights allow the Arlington fire department to prioritize high-risk zones, allocate resources effectively, and enhance overall preparedness. The goal is to shift emergency management from a reactive approach to a more predictive and preventative model.​ Targeted toward professionals in the Arlington fire department, AFRAM provides a specialized tool that supports informed, data-driven decision-making. Its features include location-based fire risk analysis, real-time notification alerts, and a user-friendly interface that enables users to filter and explore risk information by zip code, street address, and environmental conditions. While not designed for the public, AFRAM plays a vital role in optimizing emergency response strategies and fire prevention planning in Arlington.

Project Requirements

  1. Real-time fire risk prediction based on incident history, weather, and environment.
  2. Live alerts and notifications are displayed in the UI.
  3. Interactive map with filters, search, and visual overlays.
  4. Cloud deployment for remote access.
  5. Fast performance and near-instant UI updates.
  6. Scalable backend architecture.
  7. Modular design to support upgrades.
  8. Responsive layout for desktop.
  9. Color-coded fire risk tiers for easy interpretation.
  10. Search bar powered by OpenStreetMap.

Design Constraints

  • Functionality: Real-time risk scores and map updates must be accurate and reliable.
  • Usability: The Interface must be intuitive for non-technical city staff.
  • Cost/Economic: The System uses open-source libraries and free hosting tools to minimize cost.
  • Scalability: The Architecture allows for expansion beyond Arlington.
  • Maintainability: Each layer is modular and independently updatable.
  • Environmental: Enables early detection and response to fire hazards, minimizing damage.

Engineering Standards

  • Security: CORS is configured on Flask backend.
  • Web Dev: RESTful API with React (TypeScript) and Leaflet.
  • Data Standards: Normalized JSON data between layers.
  • Programming: Modular codebase and reusable components.
  • NIST Compliance: Structured model prediction and risk classification.

System Overview

AFRAM consists of four core layers:

  • Frontend (React + Leaflet): Renders map visuals, UI filters, search, and heatmaps.
  • Backend (Flask + MongoDB): Hosts prediction APIs and stores processed data.
  • Machine Learning (scikit-learn): Predicts fire risk scores based on weather and incident trends.
  • Weather API Service: Pulls real-time atmospheric conditions to enhance predictions.

Results

  • Fire risk scores are accurately predicted using regression-based modeling.
  • Heatmaps and grid overlays visualize risk tiers by color.
  • Search bar allows users to query addresses using OpenStreetMap.
  • Alerts are shown via a notification bell with real-time updates.
  • Prediction data is fetched once and reused across all visual layers.

Future Work

AFRAM met its core objective: providing an intuitive, data-driven platform to identify fire risk zones across Arlington. Our design closely aligned with the Arlington Fire Department’s vision of enabling smarter resource allocation and improving emergency preparedness through visual risk mapping.

Future development may include:

  • Build out admin dashboards for internal reporting and oversight.
  • Expanding to cover more Texas cities or other fire-prone areas.
  • Adding user accounts with saved zones or alert subscriptions.
  • Enhancing risk prediction using more granular weather and sensor data.
  • Building a mobile-first version for faster field access.

Project Files

Project Charter
System Requirements Specification
Architectural Design Specification
Detailed Design Specification
Poster

References

  1. GeoSearchControl API [https://github.com/smeijer/leaflet-geosearch]
  2. Leaflet Documentation [https://leafletjs.com]
  3. Flask-RESTful Documentation [https://flask-restful.readthedocs.io]
  4. scikit-learn Documentation [https://scikit-learn.org]

Steven McDermott

More Posts