Valeria & Nia
31 & 25 · industrial designer
- Screening threshold
- $6,750/mo
- Wants
- 2BR
Walks away if
No in-unit laundry.
Engine 1 · Demand & Mix · Brooklyn
Meet the people who will rent your building.
Prelived generates synthetic prospects for a Brooklyn site — modeled households with incomes, priorities and deal-breakers, built from who is renting in comparable neighbourhoods now and where the site is heading by the time it opens — and turns them into a unit mix you can build.
One mark per generated household · 94 West Street, Brooklyn, NY 11222
600 households · 5 segments · delivers June 2030
- Nour & Ravi36%
- Ezra24%
- Julian & Zoe17%
- Sofia & Charlotte12%
- Malik11%
10×
households generated per unit — 600 for a 60-unit building
5
segments, none of them chosen by us
2026
the year the households are modeled at — the site's own signals point toward 2030
0
people surveyed — they haven't moved yet
How it works
Build the renters first. Then design for them.
Three stages, run against public data and the signals no dataset carries.
Stage 1
Read the neighbourhood forward
Liquor licences issued, restaurants opening, bodegas closing, childcare capacity added, schools improving. We model the neighbourhood at your delivery date — not today's.
Stage 2
Generate the households
Ten per unit, one at a time. Income, what they need, what makes them walk. They cluster into segments on their own; nobody picks the categories.
Stage 3
Design against all of them
Move a slider and all 600 households re-score instantly. Set a rule you never break and the model obeys it, then tells you exactly what that rule cost.
The output
Segment share is your unit mix.
5 groups emerged from this run. Their shares, their budgets, and what each one is actually looking for.
Nour & Ravi
33 & 38 · SENIOR ENGINEER & DATA SCIENTIST · SEATTLE
35.8%
Median income$253,000
Screening threshold$6,325/mo
Ezra
31 · UX DESIGNER · WILLIAMSBURG
24.3%
Median income$150,500
Screening threshold$3,762/mo
Julian & Zoe
36 & 34 · SMALL BUSINESS OWNER & SPEECH THERAPIST · WINDSOR TERRACE
16.7%
Median income$304,500
Screening threshold$7,612/mo
Sofia & Charlotte
27 & 27 · TATTOO ARTIST & LINE COOK · BED-STUY
12.2%
Median income$175,000
Screening threshold$4,375/mo
Malik
26 · PRODUCTION ASSISTANT · ASTORIA
11.0%
Median income$80,000
Screening threshold$2,000/mo
Share × unit count is where the conversation starts, not where it ends — the Model prices every unit type against the same households and moves units accordingly. Open this run in the Model →
Three of 600
Every one of them is a person you can open and read.
Valeria & Junho
27 & 23 · program manager
- Screening threshold
- $8,550/mo
- Wants
- 1BR + alcove
Walks away if
No in-unit laundry.
Trevor & Charlotte
34 & 31 · product lead
- Screening threshold
- $6,300/mo
- Wants
- 1BR
Walks away if
No package room.
Show how Valeria & Nia’s screening threshold was built
Screening threshold · Valeria & Nia
$6,750/mo
What was observed
- ObservedHousehold incomemedian $182,909
- Source
- US Census Bureau, ACS 5-year table B19001
- Geography
- 20 census tracts across the analog neighbourhoods — Williamsburg, Long Island City, Gowanus
- Period
- acs5 2019–2023
- Retrieved
- 2026-09-16
Published income for the neighbourhoods this site's trajectory is following, read across every available vintage rather than a single year.
- ConventionScreening multiple40x gross annual income
A screening rule some owners apply, set on the property rather than measured from anything. It is not law, it does not predict any individual application's outcome, and it is not our recommendation.
- Not modeledWillingness to paynot modeled
What a household would choose to pay for a new unit. Nothing public measures it, so the model does not carry a figure for it and no number here stands in for one. Affordability is absent for the same reason: clearing a screening rule is not the same as being able to absorb the rent.
- EstimatedAnalog neighbourhoodsWilliamsburg, Long Island City, Gowanus
Chosen by hand, each with written reasoning, and the choice is visible on the run. The evidence is read from the neighbourhoods this site is becoming rather than the one outside it today — that judgement is the modeled part, and it is where a developer should push back first.
How the figure is supported
- 01ProjectedRead forward to the delivery horizon
The observations above are carried to 2026, because a building leasing then will not be leasing into today's neighbourhood. How far that projection is allowed to run is limited: past a point the trend would be doing more work than the data, and this one reached that limit.
2026
- 02EstimatedMatch to this household's type
The population is built from household types — who they are, who they live with, where they are moving from. The census sets the overall income level for the whole population; how that differs between one type and another is modeled, because no public table breaks income down that way. This household belongs to the type named below.
Dual-earner couple, remote-leaning, out of Manhattan
- 03GeneratedThis household
A synthetic household consistent with the fitted distribution — not a real person, not a survey respondent, and not a forecast of any individual. Six hundred of them stand in for the demand the building will meet.
$270,000 a year
- 04ConventionScreening threshold
What an owner applying this screening rule to this income would approve. It is a rule applied to a number, not a statement about the household: it does not say what they can comfortably afford, and it does not say what they would choose to pay. Neither of those is modeled here.
$6,750/mo
Assumptions
- Illustrative 40× gross annual income — a screening convention some owners apply, not law, and not a guarantee of approval. The multiple is a property setting.
- Neither affordability nor willingness to pay is modeled. Clearing a screening rule is not the same as being able to absorb the rent, and neither is the same as choosing to pay it.
- Income is drawn from the fitted distribution, so the population reproduces the census figures in aggregate rather than repeating them household by household.
- Figures are rounded for presentation.
What this does not establish
- The threshold is a screening convention, not a market read. It says what a rule would approve, not what anyone would rent or at what price.
- ACS median gross rent is occupied stock, including rent-stabilised units. It is the wrong base for a new building's asking rent and is deliberately not used to set one.
- Observed rent burden in these neighbourhoods is shown alongside for comparison only. It measures what sitting renters already pay under tenancies and prices set years earlier, and nothing here is derived from it.
- The differences between household types are modeled, not measured. ACS does not publish income by household archetype.
- This is a modeled household, not a person. Nobody was surveyed.
- Nothing here has been validated against a completed lease-up.
- 2 of 22 analog tracts returned no census record and contributed nothing.
Synthetic prospects, generated rather than written. No real resident is represented and nobody was surveyed — you can’t survey people who haven’t moved yet. Open the panel above to see what one screening threshold rests on — each input named, with its source, geography, period and the date it was retrieved, and marked as measured, modeled or a stated convention.
Amenities
A gate, not a price.
Modeled share of the synthetic prospects that will not rent without each one. We publish no dollar figure per amenity and no preference ranking either — nothing observable says what a household would pay for a roof deck, or how it weighs one against a bike room. What a household states outright is narrower, and it is something the model can carry.
In-unit laundry
42%n=75
Dishwasher
26%n=47
Pet amenities
17%n=31
Leave one of these out and those households are gone from the pool entirely — that is what makes it a gate. Amenities named by fewer than 25 of the synthetic prospects are left out rather than shown thin.
Why comps fail
A comp tells you what leased. It can't tell you who walked away.
It reports the sticker price
A comp records asking rent. What a landlord collects is asking rent minus the concessions it took to fill the unit, and nothing obliges anyone to update the sign.
It describes a neighbourhood that already happened
Comps are buildings leased last year to people looking then. You open in 2030, into somewhere measurably different.
It averages people into percentages
A modeled 26% of these synthetic prospects will not rent without dishwasher at all. A comp records the building that leased, not the households that never applied to it.
It can't be asked a follow-up question
Switch off in-unit laundry here and the model tells you how many households leave your pool entirely. There's nothing to run that query on in a comp set.
What it isn't
Development is art, not science. We're not arguing.
A developer who has leased six buildings knows things six hundred modelled households never will. Tell Prelived you never build studios, or that in-unit laundry is non-negotiable in your buildings, and it obeys without negotiation — it cannot return a program that breaks your rule.
What it does instead is price the rule: this one costs you three points and $4,100 a month, that one is free because the population already agrees with you. Better colours, a bigger canvas, and a note on what each choice cost. The painting is still yours.
Design partner
Ready to meet them?
Bring a Brooklyn ground-up rental, 20–150 units, with a delivery date. You get the population, the segments and the Model at no cost during the design partner phase — in exchange for an honest read on whether it's useful, wrong, or missing something.
Prelived has not been run against a completed lease-up. Nothing here is validated against an outcome. What every number does carry is a label: fitted from a named public table, a declared prior, or a stated convention. Income, household composition and remote-work share are fitted; amenity rankings and the differences between household types are priors; the screening multiple is a convention the property sets. Each household says so about itself.