Standard metrics — rent, square footage, commute time — describe the apartment. They don't describe the day. And the gap between those two things is, in my experience, the actual answer to whether a neighborhood works for you.
The question I want to spend this series on is what that gap looks like when you measure it directly:
Livability is the cumulative effect of thousands of daily frictions and affordances — and NYC is one of the most measurable cities anywhere, so let's actually measure them.
This is the first post in Measuring New York, a long-form series of eleven chapters — one per dimension of livability — with every analysis reproducible from a separate analysis repo. Today is the framing post: why this is hard, why it's worth doing, and one small map to set the tone.
"Livable" is a slippery word
Most public lists of "best places to live" are, structurally, the same trick. Take 5–10 metrics, normalize them, weight them by something — sometimes a survey, sometimes a hunch — and produce a ranking that puts a midsize college town first and reads almost identically to the list before it. The methodology is usually footnoted; the weights almost never are.
I don't think this is dishonest, exactly. I think it's the wrong question. A single ranking presumes a single reader: the same household, the same life stage, the same constraints. New York City is many cities, and most of its residents are not the median household, the median commuter, or the median anything. Asking "which neighborhood is best?" of a 9-million-person city is like asking which car is the best car. The answer is for whom, to do what, and at what cost?
So the project I'm starting isn't a ranking. It's a measurement exercise that produces:
- For each of eight dimensions, an honest map of where the friction concentrates
- For each map, a short list of comparable neighborhoods and the tradeoffs between them
- At the end of the series, three composite views by persona — not "the best neighborhood," but "the best neighborhood if you're a parent of small children with a 45-minute job-access tolerance and asthma in the household"
The eight dimensions come from a brief I wrote to myself before starting: mobility, housing, environmental quality, daily needs, public space, safety, economic opportunity, time-and-stress. They aren't mine — every urbanist writes some version of the same list — but they form the spine of what's coming.
Why NYC, and why now
A few things make this project feasible here, in a way it probably wouldn't be in most cities I've lived in:
- NYC Open Data is exceptionally good. 311 complaints, HPD violations, NYPD complaints, DOT crash data, sanitation routes, business filings, restaurant inspections — all of it queryable, most of it updated within days.
- The MTA publishes its full schedule as GTFS, which lets you compute "what jobs can someone in Far Rockaway reach in 45 minutes on a Tuesday at 8 AM" rather than guess.
- The geography is well-defined. New York has 59 Community Districts (CDs), each with a city-mandated review of quality-of-life issues. Every dataset I'll use joins back to them. They're big enough for stable rates and small enough to be meaningfully different — and they're the default unit of analysis for this whole series.
- The Department of Health publishes neighborhood-level air quality and asthma data. Most cities don't. Some chapters in this series would be impossible without it.
That last point is the one that keeps surprising me. NYC is one of the most measurable cities anywhere — not because the city is more transparent than its peers, but because the city's agencies operate at a scale where the act of measurement is itself an administrative necessity. 311 alone generates roughly 10,000 service requests a day. The dataset has 38 million records. Every one of them is a person who picked up the phone or opened an app to flag a friction in their life.
Which is a useful first thing to look at.
A first look: where do calls cluster?
For this opening post I want to give you one map — small, simple, no policy claim attached — to set the tone. The teaser below is 311 service requests per square mile across all 59 community districts over the last 30 days (April 2 through May 1, 2026). I normalized by area rather than population on purpose: this map is about where the density of friction runs hot, not where the population is bigger.
A few things to notice before we move on.
The hot band running up the spine of Manhattan is the easy story: density of people, density of mixed-use buildings, density of somebody else's behavior that you can call about. Midtown lit up hardest. So did the Lower East Side and parts of Chelsea.
The harder story is that the second-hottest area on this map is not in Manhattan. It's CD 12 in the Bronx — Wakefield and Williamsbridge — and the call mix there is different: it's heavily weighted toward heating, plumbing, and noise from neighbors rather than the commercial-noise and food-vendor profile that dominates Manhattan. Two CDs can have similar call rates and have those calls mean completely different things about daily life. We'll come back to this in Chapter 8.
The cool blue of Staten Island and the eastern edges of Queens is also worth flagging. Lower 311 density doesn't necessarily mean "fewer problems" — it often means "lower expectation that calling 311 will produce a result," or "different problems that aren't 311-shaped." This is the kind of thing that's easy to miss in a single-metric ranking.
Methodological commitments
Before I write a single chapter, here's what I'm promising myself — and you — about how this series will be put together. These aren't aesthetic preferences. They are the things I think distinguish a measurement exercise from a ranking exercise.
Community Districts are the default unit. All 59 of them. CDs are big enough that rates are stable, small enough that they correspond to actual lived neighborhoods, and they're the unit the city itself uses for quality-of-life review. When a chapter needs a finer grain — heat vulnerability, asthma, air quality — it'll drop to census tracts and say so explicitly.
No composite "livability score" before Chapter 9. None. Not in passing, not as an intuition pump. Composite scores almost always hide what they're doing, and once you read one you can't unsee it. The synthesis chapter at the end of the series will ship three different composites built around three different personas — explicitly to refuse the single-ranking trap.
Every chapter ends with "what this measure misses." Not as a footnote — as a section. The places measurement fails are at least as informative as the places it succeeds, and pretending otherwise is how you end up with a confident-sounding ranking that's wrong in the ways that matter most.
Vintages are frozen for the whole series. Every chapter uses the same ACS release (2019–2023, 5-year), the same MTA GTFS snapshot, the same Open Data cutoff. This is the only way comparisons across chapters can stay honest.
Everything is reproducible. Every chart in this series ships with a comment in its source that points back at the exact notebook cell that produced it. The analysis repo is public. If you want to run my pipelines and get my numbers, the answer should always be: make chapter-N. (Today's teaser: analyses/chapter-00.)
What's coming, in rough order
The next chapter — and the one I'm most excited to write — is Mobility and Access. The premise: in New York, the minutes between you and the rest of the city are opportunity access. The size of your world is measured in minutes-to-anywhere, not miles, and the shape of the city as seen through a 45-minute transit isochrone looks almost nothing like the shape on a road map. I want to draw that map.
After that, in order: Housing (rent burden, HPD violations, the geography of staying put), Environmental Quality (PM2.5 and asthma, finally with a tract-level drill-down), Daily Needs (15-minute walksheds for groceries, pharmacies, childcare), Public Space, Safety (which is much broader than crime statistics), Economic Opportunity, and finally Time and Stress — the signature chapter on the cumulative cognitive load of living here.
Then The Friction Map: the synthesis chapter where we earn the right to talk about composites, by talking about three of them, for three different kinds of people. And a short Methodology post to close.
If any of this is interesting, follow along. If a specific dimension matters to you and I haven't done it justice, write to me — corrections and counter-examples are exactly what this series should attract.