⚙️ How-To Building permits & new housing Maps & visualisation Notebook By

How to Measure the Apartment Share of New Housing by County

On this page

Problem statement

Is new housing in a county mostly houses or mostly apartments? The Census Bureau's permit files split every county's units by the size of the building, and the column for buildings of five or more units is the usual measure of apartment construction. One year of it is lumpy, because a single apartment block can be most of a small county's year, so this page pools 2021–2025, sets a floor on how many units a county needs, and then ranks, relates the share to density and maps it.

Quick answer

Sum five annual files by county and divide the units in five-plus buildings by all units:

five = pd.concat([permits(f"co{y}a.txt") for y in range(2021, 2026)])
c = five.groupby("GEOID")[["units", "est5_units", "est1_units"]].sum()
c["share5"] = c.est5_units / c.units
print(f"US 2021-2025: {c.est5_units.sum() / c.units.sum():.1%} of units in buildings of 5+; counties with 500+ units: {(c.units >= 500).sum():,}, median share {c[c.units >= 500].share5.median():.1%}")
US 2021-2025: 34.0% of units in buildings of 5+; counties with 500+ units: 1,272, median share 15.8%

Over 2021–2025, 34.0% of all units permitted in the US were in buildings of five or more. The typical county is far lower: among the 1,272 counties that permitted 500 or more units, the median share is 15.8%. Apartments are concentrated in a few large counties.

Histogram of the share of 2021-2025 units in buildings of five or more for counties with 500 or more units: a tall bar near zero and a long tail toward 100%, with the median county at 15.8%. Histogram of the share of 2021-2025 units in buildings of five or more for counties with 500 or more units: a tall bar near zero and a long tail toward 100%, with the median county at 15.8%.
Apartment share of new housing across 1,272 counties: the median county is 15.8%.

Pooling years and setting a floor

1. One year is lumpy

one = permits("co2025a.txt").set_index("GEOID")
one = one[one.units >= 100]
print(f"counties with 100-499 units in 2025: share 0% in {(one[one.units < 500].est5_units == 0).mean():.1%}, 50%+ in {(one[one.units < 500].est5_units / one[one.units < 500].units >= .5).mean():.1%}")
print(one[(one.units < 300) & (one.est5_bldgs == 1)].nlargest(3, "est5_units")[["name", "units", "est5_units"]].to_string(index=False))
counties with 100-499 units in 2025: share 0% in 54.9%, 50%+ in 8.1%
         name  units  est5_units
Itasca County    249         132
 Mower County    136         102
  Clay County    297          86

In 2025 alone, 54.9% of counties with 100 to 499 units permitted no five-plus units at all, and 8.1% had a share of 50% or more. In Mower County, Minnesota, a single building of 102 units was most of the year's 136. Pooling five years and requiring 500 units is what keeps one project from colouring a county.

2. Counties where most new homes are apartments

c = c.join(five.drop_duplicates("GEOID").set_index("GEOID")[["name", "state"]])
print(f"counties with 5,000+ units: {(c.units >= 5000).sum()}")
print(c[c.units >= 5000].nlargest(8, "share5")[["name", "state", "units", "share5"]].round(3).to_string())
counties with 5,000+ units: 345
                               name state  units  share5
GEOID                                                   
36061               New York County    36  17983   0.999
36005                  Bronx County    36  30773   0.988
36047                  Kings County    36  52476   0.980
06075          San Francisco County    06   7883   0.955
36081                 Queens County    36  27844   0.937
11001                Washington  DC    11  18793   0.916
25025                Suffolk County    25  14742   0.909
24510  Baltimore (Independent City)    24   6873   0.880

New York's boroughs lead: 99.9% of the units permitted in Manhattan (New York County) and 98.0% in Brooklyn (Kings County) were in five-plus buildings, then San Francisco at 95.5%. state is the two-digit code (36 New York, 06 California, 11 District of Columbia).

Horizontal bars of the most and least apartment-heavy counties with 5,000 or more units: New York, Bronx, Kings, San Francisco and Queens counties at 93.7% to 99.9%, and St. Tammany Parish, Liberty County and Midland County at 0%. Horizontal bars of the most and least apartment-heavy counties with 5,000 or more units: New York, Bronx, Kings, San Francisco and Queens counties at 93.7% to 99.9%, and St. Tammany Parish, Liberty County and Midland County at 0%.
Most and least apartment-heavy counties with 5,000 or more units, 2021-2025.

3. And where almost none are

print(c[c.units >= 5000].nsmallest(6, "share5")[["name", "state", "units", "share5"]].round(3).to_string())
                              name state  units  share5
GEOID                                                  
22103           St. Tammany Parish    22   6440   0.000
35000  New Mexico Balance of State    35  11753   0.000
48291               Liberty County    48   6420   0.000
48329               Midland County    48   5077   0.000
12017                Citrus County    12  10358   0.001
28033                DeSoto County    28   6140   0.001

Among counties that permitted 5,000 or more units, St. Tammany Parish in Louisiana and Liberty and Midland counties in Texas permitted none in five-plus buildings. New Mexico's "Balance of State" row also appears with 11,753 units and no apartments: it is not a county, and it belongs out of any county ranking.

Density, metros and the map

The data used on this page

The US Census Bureau's Building Permits Survey county files (annual 1990–2025 and year-to-date 2025–2026), its Northeast place files for 2020–2025, the Vintage 2025 county population estimates and the 2024 cartographic boundary counties, all public domain. They were downloaded on 16 October 2026 and saved on this site in one archive, because Census revises recent months as late reports arrive.

from pathlib import Path
import zipfile

import geopandas as gpd
import numpy as np
import pandas as pd
import requests

Path("data").mkdir(exist_ok=True)
# US Census Bureau Building Permits Survey: annual county files co1990a.txt-co2025a.txt, year-to-date county files
# co2501y.txt-co2512y.txt and co2601y.txt-co2608y.txt, Northeast place files ne2020a.txt-ne2025a.txt; the Vintage 2025
# county population estimates; 2024 cartographic boundary counties. Downloaded 16 October 2026, saved here unchanged.
# Live: https://www2.census.gov/econ/bps/County/co2025a.txt (and the same folder for the others),
# https://www2.census.gov/econ/bps/Place/Northeast%20Region/ne2025a.txt,
# https://www2.census.gov/programs-surveys/popest/datasets/2020-2025/counties/totals/co-est2025-alldata.csv,
# https://www2.census.gov/geo/tiger/GENZ2024/shp/cb_2024_us_county_20m.zip
ARCHIVE = "census-building-permits-2026-10-16.zip"
if not Path(f"data/{ARCHIVE}").exists():
    Path(f"data/{ARCHIVE}").write_bytes(requests.get(f"https://data.spatialworkflow.io/{ARCHIVE}", timeout=600).content)
Z = zipfile.ZipFile(f"data/{ARCHIVE}")

SIZES = ["1", "2", "34", "5"]  # buildings with 1 unit, 2 units, 3-4 units, 5 or more units
COLS = (["survey", "state", "county", "region", "division", "name"]
        + [f"{k}{s}_{m}" for k in ("est", "rep") for s in SIZES for m in ("bldgs", "units", "value")])


def permits(name):
    """One county file: two header rows and a blank line, then a row per county. est = estimate with
    imputation for places that did not report, rep = reported only."""
    d = pd.read_csv(Z.open(name), skiprows=3, header=None, names=COLS, skipinitialspace=True,
                    dtype={"survey": str, "state": str, "county": str, "name": str})
    d["survey"], d["name"] = d.survey.str.strip(), d.name.str.strip()
    d["state"], d["county"] = d.state.str.strip().str.zfill(2), d.county.str.strip().str.zfill(3)
    return d.assign(GEOID=d.state + d.county,
                    units=d[[f"est{s}_units" for s in SIZES]].sum(axis=1),
                    reported=d[[f"rep{s}_units" for s in SIZES]].sum(axis=1),
                    value=d[[f"est{s}_value" for s in SIZES]].sum(axis=1))


pop = pd.read_csv(Z.open("co-est2025-alldata.csv"), encoding="latin-1", dtype={"STATE": str, "COUNTY": str})
pop = pop[pop.SUMLEV == 50].assign(GEOID=lambda p: p.STATE + p.COUNTY)
Path("data/cb_2024_us_county_20m.zip").write_bytes(Z.read("cb_2024_us_county_20m.zip"))
counties = gpd.read_file("data/cb_2024_us_county_20m.zip")
counties = counties[~counties.STUSPS.isin(["AK", "HI", "PR"])].to_crs(5070)
y25 = permits("co2025a.txt")
print(f"2025 county file: {len(y25):,} rows, {y25.units.sum():,} units permitted; {len(pop):,} counties in the estimates")
2025 county file: 3,029 rows, 1,431,616 units permitted; 3,144 counties in the estimates

Example 1: share against population density

area = counties.set_index("GEOID").ALAND / 1e6
c = c.join(pop.set_index("GEOID").POPESTIMATE2025).join(area)
c["density"] = c.POPESTIMATE2025 / c.ALAND
d = c[c.units >= 500].copy()
d["dclass"] = pd.cut(d.density, [0, 50, 250, 1000, 1e6], labels=["<50/km²", "50-250", "250-1,000", "1,000+"])
print(d.groupby("dclass", observed=True).apply(lambda g: pd.Series({"counties": len(g), "share5": g.est5_units.sum() / g.units.sum()}), include_groups=False).round(3).to_string())
           counties  share5
dclass                     
<50/km²       567.0   0.182
50-250        490.0   0.241
250-1,000     157.0   0.430
1,000+         43.0   0.644

The share rises steadily with density: 18.2% in counties under 50 people per km², 24.1% at 50 to 250, 43.0% at 250 to 1,000 and 64.4% in the 43 counties above 1,000. ALAND from the boundary file is in square metres.

Example 2: the map

m = counties.merge(c[c.units >= 500][["share5"]].reset_index(), on="GEOID", how="left")
m["cls"] = pd.cut(m.share5, [0, .05, .15, .3, .5, 1.01], right=False)
print(m.cls.value_counts(sort=False).to_string())
ax = m.plot(column="cls", cmap="PuBu", legend=True, missing_kwds={"color": "lightgrey"}, figsize=(10, 6))
ax.set_axis_off()
cls
[0.0, 0.05)     377
[0.05, 0.15)    231
[0.15, 0.3)     297
[0.3, 0.5)      216
[0.5, 1.01)     136
Map of United States counties shaded by the share of 2021-2025 housing units in buildings of five or more, darkest in big metropolitan counties on the coasts and in Texas, Colorado and the Twin Cities; grey counties permitted fewer than 500 units. Map of United States counties shaded by the share of 2021-2025 housing units in buildings of five or more, darkest in big metropolitan counties on the coasts and in Texas, Colorado and the Twin Cities; grey counties permitted fewer than 500 units.
Apartment share of new housing by county, 2021-2025.

Of the counties with 500 or more units, 377 had a five-plus share under 5% and 136 a share of 50% or more. The apartment counties are the big metros; the grey counties permitted fewer than 500 units in five years.