The Top Ten of 2025

As we wrap up the year, we’ve collected our most read articles and most streamed podcasts of 2025. Check them out below and, as always, thank you to all our readers and listeners!

1. What Mayors Need to Know About SNAP 

   This podcast features Dr. Sara Naomi Bleich, professor at the Harvard T.H. Chan School of Public Health and former USDA Director of Nutrition Security and Health Equity. She discusses changes to the Supplemental Nutrition Assistance Program (SNAP) and shares actionable guidance for mayors and city leaders, including what city officials can do to ensure residents maintain access to essential nutrition support.

2. Transforming City Operations with StatGPT

   This paper, written by Stephen Goldsmith, introduces a new approach to performance management, which integrates advanced artificial intelligence and data technologies with traditional "stat" models to enhance capabilities and democratize insights.

3. Innovating in Atlanta: A Social Housing Model for the US

   The article explains how Atlanta is pioneering a social housing model through coordination, the strategic use of publicly owned land, and strong mayoral leadership. This had lead the city to rapidly build and preserve mixed-income, publicly owned rental housing, modeling a scalable template for other cities.

4. Data-Driven Models for Eviction Prevention

   Learn how Allegheny County’s Department of Human Services partnered with Carnegie Mellon data scientists to build and pilot a predictive model that uses integrated human services data to target rental assistance toward residents at highest risk of eviction and homelessness. This article also offers practical recommendations for other jurisdictions considering similar data-driven eviction prevention efforts, particularly in light of shrinking budgets and limited financial assistance.

5. The Power of Community-Driven Small Language Models

   This episode features MIT Professor Sarah Williams and Boston CIO Santi Garces, who discuss how GenAI can help governments and community members make sense of vast amounts of qualitative data in order to improve services and increase transparency. 

6. With Vision Zero, NYC DOT Improves Transportation Safety for All

   The article shows how New York City’s Vision Zero program uses cross‑agency data, geospatial analysis, and evidence-based street redesigns (like targeted turn calming and “hardened daylighting”) to sharply reduce traffic fatalities and injuries for pedestrians, cyclists, and drivers while engaging communities in safer, more bike‑ and pedestrian‑friendly streets.

7. Seeking Better Parking Data, Boston Breaks Up its Giant Contract

   Read how the city of Boston restructured a massive, decades-old parking contract into four interoperable modules, after years of being “locked in” with a single vendor, to gain better data, modern technology, and long‑term flexibility, offering a procurement model other cities can use to avoid vendor dependence and improve curb management.

8. StatGPT: The Future of City Performance

   This episode features Santi Garces and host Stephen Goldsmith discussing how GenAI tools can democratize and supercharge traditional city performance management (“stat” programs) by giving more public employees intuitive, conversational access to data while helping them more easily analyze operations, regulations, and contracts.

9. Harnessing AI for Trust and Transformation

   Hear from Miguel Carrasco, Global Leader for Boston Consulting Group’s Center for Digital Government, who explains how governments around the world are using generative and agentic AI to cut bureaucracy, improve service delivery, and enhance data-driven decision-making. He also touches on the relationship between high-quality digital services and trust in government.

10. Generative AI and the Possibility Government

    This podcast episode, with Santi Garces and Professor Mitch Weiss, explores how GenAI can transform city performance management and service delivery by empowering more public employees and community groups to use data, synthetic agents, and AI-assisted coding for problem identification, policy testing, and solution design.