As the climate crisis escalates, the old ways of managing city infrastructure—siloed departments, static plans, and reactive maintenance—simply won’t cut it. The wildfires, floods, and heatwaves hitting Southern California are not isolated incidents; they’re symptoms of deeper systemic fragility. If we’re going to build true resilience, we need to look at our cities as whole systems—where housing, mobility, land use, water, and energy are deeply interconnected.

The future of infrastructure isn't just smarter—it’s more integrated. And that integration starts with better visibility into the systems we already have.

In the post below, my cofounder Varun shares how we prototyped a low-cost, city-scale street survey in Syracuse—and how it hinted at something bigger: the power of continuous sensing, intelligent routing, and shared infrastructure data to inform smarter decisions. From waste collection routes to methane leak detection, there’s a throughline here that applies across every facet of urban management.

Efforts like CAMP4W are beginning to stitch together these threads—offering adaptive, region-wide planning frameworks that can be "open sourced" across agencies and sectors. But we need to go further. The tools are here. The data is cheap. The need is urgent. It's radically common sense to not artificially isolate utility maintenance by water or gas or power. Leveraging low cost sensors for integrated asset management and empowering adaptive approaches to managing this mission critical infrastructure is radically common sense.

It’s time to stop treating infrastructure as a patchwork of projects and start managing it as a dynamic, interconnected ecosystem.

In the previous post, I discussed how we implemented a low-cost citywide street quality survey in Syracuse.

An early challenge we faced was to develop easy to follow route plans for city workers to perform "complete" surveys across the city. At the time, we relied on the Syracuse public works department to follow their historical resurfacing plans and we just winged it.

That approach mostly worked but it left me wanting to "nerd harder" to scale more easily.

The academic literature references the Chinese Postman Problem or Route Inspection Problem, attributed to the Chinese Mathematician Meigu Guan on how to efficiently route postal workers to walk every street at least once and return to the starting point with minimal repetition.

During Mao's Great Leap Forward of 1958–1960 that transformed it from an agrarian society into an industrialized one, Chinese academics were "encouraged" to work on practical problems to serve the nation.

Guan's solution was a generalization of Leonhard Euler's famous Seven Bridges of Königsberg problem of 1736 that required creating routes that crossed each bridge exactly once. Guan published his work in 1960, and his paper was translated into English in 1962.

The solution spread widely and was used to plan sanitation and snow ploughing routes. In the early seventies, Cleveland, Ohio used it halve the cost of its waste collection program!

Removing waste from waste collection is so deliciously meta

In this sim (generated using Gemini 2.5 Pro) I demonstrate a naive version of this.

We assume that a surveyor starts from a central location and can drive 5 hours each day at an average 20 mph. This works out to 100 miles a day. We segment the city into a uniform grid. Each vehicle visits their designated square and tries to visit every street segment with minimal repetition.

We programmed the raspberry pi to simply snap an image every second and when the vehicle came to a stop for more than a minute, transmit the stored images to cloud storage.

During my stint at Urbint (2019-2022), I learnt that this is also how Gas utilities perform their federally mandated system integrity surveys. They chunk up their service territory into a grid and assign a surveyor who carries a methane detector to identify signifcant leaks.

In 9 days (2 work weeks) 3 vehicles have surveyed 2500 miles across Syracuse.

Equipping each with a simple internet connected accelerometer + camera pointed at the street level or methane detecting sensors could continuously at an interval of 1Hz would lead to a few gigs of image data that cost pennies to host on the cloud.

If every municipal department in the world were capable of collecting and acting on continuous data collection like this, we could extend the life of our public assets and serve urban needs in frictionless ways.

A new world is possible. An alternative to the dog eat dog world dominating DC.