LA County open-sources analytics model
- Los Angeles County’s Analytics Center of Excellence used a March 6 SCALE talk to formalize an open-source transparency model for public-sector analytics. - The concrete stack is Python, GitHub, and the county’s Justice Hub, with standardized charge-classification methods already powering rearrest and court-delay analyses. - It matters because LA County is turning transparency from a values statement into an operating model other governments could copy.
Government analytics is usually a black box. A county publishes a dashboard, maybe a policy memo, and the public is supposed to trust that the numbers mean what officials say they mean. Los Angeles County is trying a different move. Its Analytics Center of Excellence — inside the Chief Executive Office — is laying out a public, open-source style for how official analysis gets built, shared, and checked. That matters because the fight over public-facing analytics is no longer just about accuracy. It is about whether anyone outside the building can see the logic at all. ### What actually changed? The immediate news is a public presentation of that model. At the SCALE 23x conference on March 6, 2026, LA County staffers George Miranda and Fei Wu framed the county’s “Transparency Stack” as a repeatable approach for public-facing analytics and trust. The key point was not just that the county uses open-source tools. It was that the county is treating shared methodology as part of the product, not a side document. (socallinuxexpo.org) ### Who is doing this? This is coming from the county CEO’s Analytics Center of Excellence, or ACE. Wu is a lead research scientist in the Chief Executive Office, and Miranda is a senior data scientist on the ACE team. Their remit is broad — county departments, justice reform, poverty work, and other public-service analysis — so this is not a one-off lab demo. It is being pitched as a working county practice. (socallinuxexpo.org) ### What is in the “stack”? Basically, three layers. Python for the analysis itself. GitHub for sharing code and methodology with partners. And the county’s Justice Hub for public data dissemination. That sounds mundane, but the point is the combination: the public-facing number is supposed to sit on top of a visible method, not a hidden spreadsheet passed around by email. (socallinuxexpo.org) ### Why does open-source matter here? Because reproducibility is the real trust test. If a county says rearrest rates changed, or court delays worsened, the public and partner agencies need to know how those categories were defined and processed. ACE’s pitch is that open-source workflows make interpretation less squishy. People can inspect the logic, challenge assumptions, and rerun the work instead of arguing only about the final chart. (socallinuxexpo.org) ### What has LA County already used it for? Not just prototypes. The presentation says ACE built a standardized methodology for categorizing and describing criminal charges, and that standardized work now underpins reporting on rearrest rates for people served by the Office of Diversion and Reentry. It also supports analysis of delays in the court hearing process. In other words, this is already attached to sensitive justice-system questions where definitions really matter. (socallinuxexpo.org) ### Why is governance part of the story? Because transparency without decision rules is still messy. LA County already has a federated IT governance framework that is supposed to make technology decisions more transparent, comprehensive, and risk-aware across departments. The analytics model fits inside that bigger county posture — local departments can move, but enterprise governance still sets guardrails around standards, interoperability, and risk mitigation. (socallinuxexpo.org) ### Is this really about cameras? Not exactly — at least not in the public materials tied to this presentation. The stronger documented claim is broader: public-facing analytics, especially in criminal-justice settings, should expose methods and definitions. You can infer how that would matter for camera-based systems, where model behavior, retention rules, and escalation paths are politically loaded. But the source material here is more squarely about analytics governance than a specific county camera rollout. (ceo.lacounty.gov) ### What is the bigger takeaway? LA County is trying to make transparency operational. Not “trust us,” but “here is the code, here is the method, here is the data pathway.” If that sticks, the interesting part is not the software. It is the precedent — a very large local government treating explainability as infrastructure, not PR. (socallinuxexpo.org)
