Staff Reports
When a hurricane approaches Florida, the state prioritizes rapidly deploying resources to high-risk areas.
Emergency supplies, first responder teams, and evacuation plans are coordinated based on real-time or near-real-time data flowing into the State Emergency Operations Center. Following the hurricane, the focus shifts to recovery, including processing thousands of invoices for supplies and services to ensure communities receive aid. Leaders at all levels rely on accurate and efficient data to make these critical decisions.
The Florida Division of Emergency Management (FDEM) is modernizing its operations to enhance data usage, enabling a swifter emergency response while responsibly managing taxpayer funds. This includes developing a robust data collection and analysis system and a machine-learning tool for detecting unusual invoices.
“Slalom is proud to partner with such an innovative agency that directly addresses the challenges faced by our citizens, businesses, and communities in the aftermath of a hurricane or other emergencies,” said Beau Williamson, Florida General Manager of Slalom, a technology consulting firm collaborating with FDEM since Hurricane Ian.
The data platform, built on cloud infrastructure, underpins the analytics FDEM needs for decision-making and operational efficiency. Key components include:
-
Centralized data platform: Integrates and securely stores information from the Division of Emergency Management Solution (DEMES) and other sources.
-
Data governance strategy: Establishes rules and guidelines to ensure data clarity, accuracy, and accessibility.
-
Advanced analytics: Provides end-user tools that unlock significant value, including finance and procurement dashboards, invoice anomaly detection, and GenAI for disaster declarations.
Invoice anomaly detection
The invoice anomaly detector is a vital part of FDEM’s data platform, designed to identify and mitigate financial risks by spotting anomalies in the invoice payment process. The detector:
-
Identifies and interprets various types of invoice discrepancies with over 99% accuracy.
-
Upholds the state’s commitment to financial integrity by quickly detecting and addressing unusual invoices.
-
Streamlines the invoice process for faster resolution and improved resource allocation.
In its first 30 days of deployment, the anomaly detector identified three significant invoice irregularities totaling nearly $600,000. It analyzes historical invoice data to identify trends and deviations, training a machine learning model to differentiate between normal and abnormal invoice behavior. When an anomaly is detected, the system sends notifications for further review. Ongoing enhancements aim to optimize financial operations and advance invoice anomaly detection.
Transformative Insights
Data is crucial for leaders tasked with protecting and serving Floridians before, during, and after a disaster. By leveraging advanced analytics and machine learning, FDEM has transitioned from manual information collection to a modern, data-driven approach. This shift significantly enhances their ability to analyze and act on transformative insights rather than spending time sorting through unstructured data from multiple sources.
Florida is at the forefront of technology, emergency management, and financial integrity. These advancements empower FDEM’s workforce to effectively fulfill its mission and support Florida’s economy after a disaster.
