NYC Capital Projects · Open Data

One prompt,
one report.

Four interactive analyses of New York City's capital portfolio — each generated from a single prompt to the OD-CPD MCP server, with the domain rules (PID↔FMS many-to-many, role-aware agency attribution, signed variance) already baked in.

Source · NYC Open Data 4 Socrata sets → DuckDB Served over MCP Snapshots · 202601–202605
Four sheets below
SCHEDULEAGENCIESBUDGET
Dataset field guide

NYC capital projects overview

a first look at what the data covers

A nine-slide introduction to schedule and budget units, reporting coverage, lifecycle mix, budget and spend, agency roles, and 22 category buckets.

Open presentation →
Drawingdataset_overview
Period202605
SourceNYC OpenData
Topology

Schedule ↔ Budget topology

the many-to-many anatomy of the portfolio

Fan-out rings, a bipartite diagram, the “tangled few” outliers — hover to see the real schedules and budget lines — an agency scatter, and a budget concentration curve.

Open report →
Drawingpid_fms_budget
Period202601
SourceNYC OpenData
Parks · $50M+

Parks projects over $50M

every big build, and what funds it

23 budget lines; hover any to reveal its linked schedules. Quietly applies the category taxonomy, so the $1.9B “Park Pedestrian Bridges” route to Bridges, not Parks.

Open report →
Drawingparks_over_50m
Period202601
SourceNYC OpenData
Change monitor

Budget & schedule change monitor

what moved this period, by agency

KPIs, a trend chart, a sortable watchlist, and a per-project popup with schedule-variance bars and a stacked budget-vs-spend chart.

Open report →
Drawingchange_monitor
Period202601
SourceNYC OpenData