New York Police
Department

The Challenge

NYPD commanding officers and their operations staff were managing precinct activity through a mix of radio logs, manually updated whiteboards, and a legacy CAD (Computer-Aided Dispatch) system that exposed data only to the Communications Division — not to precinct command staff who needed situational awareness most.

Shift commanders had no consolidated view of active incidents, officer deployment status, or resource availability across their borough. Cross-precinct coordination during major events required phone calls to neighboring precincts. After-action reporting was done manually, compiled from multiple sources into Word documents.

Any system built for this environment had to meet uncompromising standards: no single point of failure, encrypted data at every layer, role-based access with no data bleed between precincts, and hardware performance optimized for the older workstations deployed in precincts throughout the five boroughs.

The Solution

We embedded with NYPD's IT division and a cross-section of precinct commanders during a two-week discovery phase. The goal was to understand what decisions commanding officers make in the first 10 minutes of an incident — and what data they needed, in what format, to make those decisions faster.

The dashboard was designed around a hierarchy of views: the Precinct View (what a commanding officer sees for their precinct), the Borough View (what an Inspector sees across multiple precincts), and the Operations Center View (what the citywide operations team sees). Each view surfaces different aggregations of the same underlying data, with permissions enforced at the API layer.

Security architecture was designed and reviewed by an independent third-party before a single line of code was written. The stack was chosen to minimize attack surface: no third-party analytics, no client-side data persistence, and all API calls mutually authenticated via certificate pinning.

What We Built

• Incident Live Feed — real-time stream of active CAD incidents within the precinct's jurisdiction, with incident type, priority, assigned units, and elapsed time. Auto-refreshes every 15 seconds without a page reload.
• Officer Deployment Map — a borough-level map view showing the last-known position of radio-equipped patrol units (from AVL data), color-coded by availability status. Commanding officers can see gaps in coverage at a glance.
• Shift Management Panel — structured view of the day's tours: who is on duty, what assignment they hold, when they go off, and any overtime or administrative assignments. Integrated with the payroll and scheduling system.
• Cross-Precinct Resource Request — a structured workflow for requesting mutual aid — additional officers, specialized units, or equipment — from neighboring precincts or borough command, replacing the phone call workflow with a trackable digital request.
• After-Action Report Generator — automated incident summary generation from CAD data, pulling timeline, units involved, incident disposition, and sector information into a formatted report template. Reduces report preparation time from hours to minutes.
• Role-Based Access Control — five distinct access tiers (Officer, Supervisor, Commanding Officer, Inspector, Operations Center) with data scopes enforced at the API layer. A supervisor cannot see data outside their precinct. Access logs are immutable and retained for 7 years.
• Offline Degradation Mode — if network connectivity is lost (a real scenario in underground or structurally complex environments), the dashboard serves a cached read-only view of the last-known precinct state from local service worker storage.

Results