Instead of engineers pulling data from different silos, the Oklahoma Department of Transportation (ODOT) has begun moving its inspection histories, condition reports and other records into Google’s BigQuery data warehouse to eventually coalesce all bridge maintenance data and make it available through dashboards and predictive models. The goal: anticipate problems before they become emergencies.
ODOT staff won’t just be utilizing the BigQuery warehouse to centralize data, but also Google Cloud’s Dataplex Universal Catalog to create a specialized business glossary for the ODOT staff and an analytics hub for secure data sharing.
While the implementation is still a work in progress, ODOT leaders are knee deep in the process: Data governance foundations are set, some bridge data has been migrated, divisions are being onboarded, and training is now underway on the Dataplex Universal Catalog.
The next milestone is integrating AI tools like Gemini in Looker, which would allow staff to eventually ask plain-language questions — for example, “Which bridges in Woods County exhibit the highest wear indicators?” — and generate insights in seconds. Google Public Sector’s Chris Hein shared those kinds of use cases as possibilities the system could support, though ODOT is still working toward fully realizing that vision.
“It standardizes the source of truth for the data sets,” Lance Underwood, ODOT enterprise systems services director, said. “It makes the analysis quicker because we don’t have to question which piece of data is accurate.” This clarity, he said, not only speeds up internal decision-making but helps the department respond more efficiently to requests from legislators and the public.
The initiative began last year, when ODOT started migrating bridge data into the Google Cloud Platform (GCP) and setting up formal governance practices. The challenge, familiar to many state DOTs, was decades of information collected across departments in inconsistent formats.
Justin Hernandez, ODOT director of design, described the shift as cultural and technical.
“Over the years, different design and production branches collected and stored data in their own ways, making it difficult for other parts of the department to access and use it,” he said. “By formalizing it in one place, anyone across the department can view it with confidence, knowing it’s accurate and has gone through a governance process.”
Future plans for the ODOT and Google collaboration include expanding the department’s AI-driven approach beyond bridges. Underwood said the goal is eventually to include “all of our production data within the GCP platform and have all of those AI data models touching all of that data so that we can ask even more insightful questions.” That could encompass crash data, roadway condition reports or even construction management information, broadening the scope of predictive insight across the agency.
Hernandez painted a picture of untapped potential within the organization, too, and said the real challenge has never been a shortage of information, but rather, figuring out how to apply it.
“We have mountains of data. We’re not terribly efficient at being able to use it,” he said. “I’m excited that this will open that up to a broader audience, and we can have confidence that the data they’re viewing is usable.”
