Enhancing emergency readiness and response with data mesh

By developing a platform capable of enabling secure data sharing across federal, state, and local partners, this modernized data hub helps the agency accelerate decision-making, strengthen oversight, and transform disaster readiness nationwide.

Challenge

For years, the client agency and its state and local partners relied on an isolated enterprise data warehouse (EDW) to guide emergency preparedness and response. During disasters, on-the-ground teams couldn’t access or share data in real time, limiting the agency’s ability to generate the insights needed for critical decisions. After disasters, the EDW made it difficult to identify and investigate fraudulent use of federal funds.

As a result, support for people and businesses recovering from disasters was often delayed—and taxpayer funds were put at risk.

Solution

We partnered with the agency to modernize its legacy EDW using Databricks, creating a scalable data lakehouse. The platform is one of the first federal implementations of an agency-wide data mesh—combining decentralized data ownership with strong centralized governance—and is built on a DevSecOps-driven architecture that improves efficiency and enables rapid deployment of analytical tools.

With this foundation in place, teams can act on data in real time. A built-in AI chatbot analyzes live disaster data to support faster, more informed decision-making during disaster declarations. After emergencies, users can apply advanced analytics to detect fraud in assistance applications and flag potential cases for investigation. Leaders also have access to integrated dashboards that bring together data from multiple systems to support projections and ad hoc reporting.

Results

Through this partnership, we’ve established a scalable model for coordination across federal, state, and local partners, with each component operating on a shared data fabric.

Among early regional adopters, usage increased by 40% in 2024–2025—improving analytics, coordination, and incident management across major disasters. Reusable datasets now support fraud detection in Individual Assistance applications, reducing analyst workloads by 50% while helping meet congressional reporting requirements.