1) Building an MCP Server for AI-Driven Analysis of NYC 311 Data // Nicholas Taub
This lightning talk features an MCP server that exposes NYC 311 Call Center data to AI assistants like Claude through tools, context and resources. The server offers three core tools — Retrieve (hybrid keyword/vector search, Analyze (aggregation and theme extraction) and Annotate (add record comments). The stack is Python-based using FastMCP, Supabase for storage and Google's Gemini for text embeddings.
AI & Software Engineer Nicholas Taub will demo the architecture showing how LLMs can infer and relay relevant records between tools. The server's tool descriptions and docstrings provide explicit usage instructions to the LLM - e.g. tool ordering, query parameters and workflow constraints - enabling the reasoning model to orchestrate tasks with minimal human intervention. Resources are optional supplementary context for more refined inputs.
This session is designed for analysts, engineers and enthusiasts interested in learning how MCP can enhance their LLM experience.
2) Civic Engagement for the Digital Age // Devin Neal, Kris Turkal
Join software engineers Devin and Kris for a live demo of Civic, a mobile app that makes it easy for anyone to find and contact their elected officials at every level of government. We'll walk through the app's core features, starting with an interactive map that visualizes city council, state legislative, and congressional districts together with data on their representatives. Then we’ll showcase the "My Representatives" page which shows your local, state, and federal officials based on your location. Attendees will see how Civic pulls together district boundaries, government directories, and other open data sources to create a rich view into our government through clean mobile interface.
This session is ideal for anyone interested in civic technology, open government data, or building data-intensive mobile applications. Whether you're a developer curious about geospatial data, a policy advocate looking for tools to connect constituents with their representatives, or simply someone who wants to know who represents you, you'll walk away with something useful. Attendees are encouraged to download the app and follow along on their own devices during the presentation.
3) Engagic: From Council Agenda to Neighborhood Input // Iban Sadowski
In 2020 a group of Palo Alto residents tried to hold their city council accountable and improve information accessibility, by hand. They tracked meetings, read and summarized agendas, and organized neighbors to show up and speak. We burned out in 4 months.
Engagic is the tool that would've saved them. It's an open source platform that automatically tracks council meetings across 100+ US cities, summarizes agenda items using LLMs, and surfaces voting records and committee decisions for elected officials. The data is freely available; the code is on GitHub.
This lightning talk will demo engagic using live NYC council data: what a meeting looks like before processing (dozens of attachments, hundreds of pages) and after (plain-language summaries, topic filters, participation links). The talk will also show the deliberation feature, which lets residents provide structured feedback on legislative items and see where their neighbors stand.
Ideal for: civic technologists, journalists covering local government, librarians, anyone who's ever tried to figure out what their city council is actually voting on.
4) NYC Community Boards Dashboard // Tina Zeng
The NYC Community Boards Dashboard (community-boards.nyc) is an interactive tool designed by BetaNYC Associate Board member Tina Zeng to support transparency, analyses, and comparative evaluation of Community Boards (CB). It uses CB digital records, Demographic Reports, District Needs Assessment to provide an overview of how community boards reflect and respond to their neighborhoods. The dashboard aims to measure civic participation, discourse quality, responsiveness, and accessibility in order to empower residents, advocates, researchers, and policymakers to explore patterns over time and gain insight into equity, engagement, and local decision-making.
The session will begin with a walkthrough of the dashboard, then present the development and analysis processes, including methodological and ethical considerations. Followed by a Q&A and feedback session, inviting participants to critically engage with the tool, and explore how similar approaches could be applied to other local institutions (this project was inspired by previous work out of School of Data like citymeetings.nyc).
Join if you’re also curious about building with AI from scratch—using natural language processing, large language models, and vibe-coding!!
