Use of Air Quality Sensor Network Data for Real-time Pollution-Aware POI Suggestion

TL;DR

This work addresses the need for privacy-preserving, real-time, pollution-aware POI recommendations in urban environments. It presents AirSense-R, comprising a Prediction Engine and an AirTown Recommendation Engine that fuse AirSENCE sensor data from Bari and Cork with user preferences, using FBProphet-like forecasting, anomaly detection, and radial basis interpolation for sparse data. Personalization is achieved through Matrix Factorization trained via Federated Learning and a privacy-preserving re-ranking score that blends CF preferences with current AQI as $S = \alpha \cdot S_{\\mathrm{MF}} + (1-\\alpha) \, S_{\\mathrm{AQI}}$ with $\\alpha \\in [0,1]$, within a ~1 km radius. Demonstrations on real and synthetic data show federated learning yields lower median absolute errors and that the system can adapt recommendations to individual health needs while preserving user privacy, highlighting feasibility for health-conscious urban navigation. The work contributes a concrete, privacy-aware framework for real-time POI suggestions that respond to dynamic air quality conditions, with potential for extension to richer environmental context and real-world evaluations.

Abstract

This demo paper introduces AirSense-R, a privacy-preserving mobile application that delivers real-time, pollution-aware recommendations for urban points of interest (POIs). By merging live air quality data from AirSENCE sensor networks in Bari (Italy) and Cork (Ireland) with user preferences, the system enables health-conscious decision-making. It employs collaborative filtering for personalization, federated learning for privacy, and a prediction engine to detect anomalies and interpolate sparse sensor data. The proposed solution adapts dynamically to urban air quality while safeguarding user privacy. The code and demonstration video are available at https://github.com/AirtownApp/Airtown-Application.git.

Figures (6)

• Figure 1: AirTown integrates real-time data from sensors installed in Bari and Cork, with user preferences to provide personalized, pollution-aware POI recommendations.
• Figure 2: The architecture of the proposed system.
• Figure 3: Weekly diurnal trends of NO2 for six AirSENCE devices in Bari, Italy 2023.
• Figure 4: NO readings for six AirSENCE devices in Cork, Ireland on January 29, 2024.
• Figure 5: NO pattern prediction using FBProphet and AirSENCE device 161 in Bari, Italy.