How the Georgia Data and Analytics Center uses AWS cloud services to drive improved outcomes for constituents

For decades, state agencies in Georgia collected enormous amounts of operational data, but most of it stayed locked inside individual departments. Analysts manually downloaded files and built Excel workbooks to track spending across programs. Agencies that needed information from other departments faced a difficult process, and executing a single data sharing agreement (DSA) could take years. Across the state government, variations of these challenges played out in every department.

In 2019, the Georgia Legislature established the Georgia Data Analytics Center (GDAC) within the Office of Planning and Budget (OPB) to break down those barriers. Housed in the Governor’s Office of Planning and Budget, GDAC serves as a central data team, helping agencies organize, share, and analyze cross-government data through a secure, governed environment. Today, working with Amazon Web Services (AWS), GDAC supports over 35 agencies, turning siloed information into the cross-agency insights Georgia’s policymakers need to better serve constituents.

From siloed data to a stalled program

State agencies had been collecting data for decades, but each one focused on its own core systems, compliance requirements, and legislative mandates. Cross-agency questions went unanswered because staff didn’t have the time, tools, or access to connect data across departments. “For example, the Department of Education could have all their students’ educational data, but they didn’t have access to their students’ correctional, healthcare, or foster records,” explained Kanti Chalasani, division director of GDAC.

The legislature created GDAC to address this, but the program’s first iteration, a platform-as-a-service (PaaS) model, struggled to gain traction. When Chalasani joined the team in late 2020, there was no working cloud environment, no production use cases, and only two DSAs in place. “The program was failing when I joined,” Chalasani recalled. “Performance was horrible. Not even a single use case was in production.”

To turn things around, GDAC needed a secure, scalable analytics environment capable of supporting sensitive government and health data, along with a cloud collaborator that could help the team build trust with agencies cautious about sharing their information.

Building a foundation of trust and security on AWS

GDAC evaluated three cloud environments. Because the data would include health and human services records, the team prioritized compliance capabilities, granular logging, and full observability into who accessed which dataset and when. No one on the team had extensive cloud experience, so the environment also had to support rapid learning alongside implementation.

AWS met all those requirements. The Georgia Department of Revenue had already built a secure AWS environment, giving GDAC a proven model to follow within the state government. Throughout the project’s development, AWS provided comprehensive support through the AWS Data Lab program, while GDAC’s dedicated team remained focused on learning and delivering. This approach built rapid skill development in parallel with developing production solutions, supported by ongoing technical guidance from AWS account teams.

With that, the engagement moved fast. Starting March 1, 2021, the GDAC team worked with the AWS Data Lab team to build a cloud analytics solution from scratch. By July 1, they had launched 10 use cases, and AWS Control Tower was established by October to centralize governance. Management transferred to the Georgia Technology Authority (GTA) in November, bringing on a chief information security officer and establishing formal security policies and audit-readiness procedures. Altogether, GDAC teams’ commitment, hands-on implementation support through Data Lab, security best practices drawn from existing state deployments, and access to AWS analytics and governance services helped GDAC to move from concept to production in those four months—while building the internal expertise to sustain it.

Trust also shaped every decision in that first year. For example, Chalasani’s team accepted only low-risk, public datasets while the secure environment matured. “We were living in a tent while we were building the building,” she said. By the time the secure environment was ready for sensitive data, GDAC had already proven to the agencies that it could be trusted to handle it.

How the environment works

GDAC now operates in two primary environments. The first is the GDAC data lakehouse, which uses Medallion architecture to store and curate data from across the state government. Agencies send data through AWS Transfer Family (SFTP), APIs, or manual uploads, and GDAC ingests and curates it using AWS Glue and AWS Lambda. From there, agencies receive analytics through Tableau and Power BI, or they can analyze the curated datasets directly with their own tools.

The second environment is the All-Payer Claims Database (APCD), built on the Observational Health Data Sciences and Informatics (OHDSI) open-source stack. A state vendor collects healthcare claims data from entities offering health insurance, pharmacy, and/or medical benefits across health providers and facilities in Georgia. GDAC loads it into the data lake and then into Amazon Redshift, runs quality checks, and transforms it into the OHDSI common data model. Researchers at the Georgia Tech Research Institute access the data via Amazon WorkSpaces using Jupyter notebooks, RStudio, Tableau, and SQL interfaces.

A broad set of AWS services supports the work across both environments. Amazon Athena reduces turnaround time from data collection to consumption, and Amazon DynamoDB powers file tracking and auditing. Amazon GuardDuty and AWS Secrets Manager support security, while AWS CloudTrail, Amazon CloudWatch, and AWS Config handle auditing and monitoring. AWS IAM Identity Center provides federated single sign-on through GTA’s central Active Directory.

GDAC operates on a use-case submission model where agencies submit questions, and GDAC identifies relevant datasets, builds pipelines, and develops dashboards through an iterative process. The team positions itself as an extension of what agencies can already do, not a replacement for them. “We are only extending your capabilities,” Chalasani said.

Built on AWS with centralized governance from AWS Control Tower and data cataloging from AWS Glue, the platform also provides the foundation to layer in AI for automated analysis, workflow automation, and anomaly detection.

Delivering results across Georgia

In the nearly five years since that first July 2021 launch, AWS has continued to support GDAC as the program scaled. The GDAC team now maintains more than 70 dashboards, split across 31 public, 30 internal, and 10 secured, and has collected more than 100 DSAs, including one of the first Medicaid DSAs of its kind in the nation. A DSA library and global data-sharing templates maintained by GDAC have directly supported the legislative process for efforts such as Senate Resolution 237 on teacher education workforce development.

That infrastructure is helping policymakers across Georgia make decisions that were previously out of reach, including:

1. Salary benchmarking across agencies to inform competitive compensation decisions
2. Open purchase order visibility, surfacing years’ worth of unresolved POs, helping to identify funds eligible for return to the state treasury
3. Healthcare enrollment analytics that give administrators the visibility they need to improve plan management
4. Education funding analysis through a rule-based engine that supports what-if policy modeling
5. Statewide health research using claims data, helping researchers uncover trends and inform better health policy

Use cases like the education funding engine, which models billions of dollars in state funding, would have been difficult to imagine before GDAC. “Things that were not possible, that nobody could even imagine, we can solve now,” Chalasani said. That track record has further solidified trust across the state government. Agencies that once lacked the capacity for complex cross-agency work now bring their most difficult problems to GDAC and get answers.

Expanding capabilities and advising other governments

GDAC will soon house historic PeopleSoft data as the state transitions to Workday, with governed, role-level security so each agency accesses only its own records. The team is also collaborating with the state’s chief artificial intelligence (AI) officer, experimenting with AI tools on public datasets, using AI to support coding and script generation, and exploring anomaly detection for financial data such as duplicate voucher payments.

Chalasani’s advice for other state and local governments considering a similar approach: start with the cloud, start small, and build iteratively. “Pick a use case that covers the whole cross-section of the solution and implement the breadth of it through that use case,” she said. “If we fail, let’s fail fast. If we succeed, let’s keep going. If we get a roadblock, let’s find a diversion.”

To go deeper into the program’s results and approach, visit GDAC’s website and annual reports. To explore how AWS can help your government build its own secure, scalable data analytics capabilities in the cloud, contact the AWS Public Sector team today.

Read related stories on the AWS Public Sector Blog:

• Driving public sector innovation: Building a data lakehouse and analytics platform on AWS for public housing
• Anne Arundel County integrates generative AI into case management to empower staff and improve citizen services
• 4 steps to build a data strategy for managing performance in the public sector
• North Carolina Division of Employment Security modernizes customer services with generative AI on AWS