Emergency Incident Detection from Crowdsourced Waze Data using Bayesian Information Fusion

TL;DR

This paper tackles proactive emergency incident detection by leveraging noisy crowdsourced Waze reports and validating against official E-TRIMS data. It introduces a Bayesian information fusion framework that models report reliability and fuses across space and time, enabling sequential updating of incident probabilities with localization in regions. The authors demonstrate improved performance over baselines on Nashville data, achieving higher F1 and AUC, and show the approach can detect incidents several minutes before official reporting (average lead time about 5.92 minutes). The framework is extensible to incorporate additional noisy data sources to support faster and more reliable emergency response.

Abstract

The number of emergencies have increased over the years with the growth in urbanization. This pattern has overwhelmed the emergency services with limited resources and demands the optimization of response processes. It is partly due to traditional `reactive' approach of emergency services to collect data about incidents, where a source initiates a call to the emergency number (e.g., 911 in U.S.), delaying and limiting the potentially optimal response. Crowdsourcing platforms such as Waze provides an opportunity to develop a rapid, `proactive' approach to collect data about incidents through crowd-generated observational reports. However, the reliability of reporting sources and spatio-temporal uncertainty of the reported incidents challenge the design of such a proactive approach. Thus, this paper presents a novel method for emergency incident detection using noisy crowdsourced Waze data. We propose a principled computational framework based on Bayesian theory to model the uncertainty in the reliability of crowd-generated reports and their integration across space and time to detect incidents. Extensive experiments using data collected from Waze and the official reported incidents in Nashville, Tenessee in the U.S. show our method can outperform strong baselines for both F1-score and AUC. The application of this work provides an extensible framework to incorporate different noisy data sources for proactive incident detection to improve and optimize emergency response operations in our communities.

Figures (8)

• Figure 1: Figure shows the varied distribution of crowdsourced Waze reports (blue) about incidents (red) in a region at two different times, illustrating the information fusion challenge for crowdsourced data across space and time.
• Figure 2: Similar hourly trend of Waze reporting and accident records in the official E-TRIMS system.
• Figure 3:
• Figure 4: Distribution of time interval (min) of clusters. Colors indicate the value of $\epsilon$.
• Figure 5: Variation of performance with incident interval.