About the data

A project of Ntontan RE

Ntontan RE develops spaces that bring people together — rooted in sustainability, culture, and connection. This atlas is part of that work: a clear, honest read of every Des Moines neighborhood, so that development and investment can start from how places actually are.

It assembles property, demographic, economic, and amenity data into a single neighborhood-level view of all 51 neighborhoods. The goal is a meaningful, honest picture — not a leaderboard. Everything below describes how the numbers are built so you can judge them for yourself; where a choice could have gone another way, the choice is stated.

The unit: neighborhoods, assigned by location

Every parcel and record is placed into a neighborhood by a spatial join — matching on actual location rather than on neighborhood names. Name-matching produced systematic misassignment; location-based joins resolved it, and that’s the pattern used throughout the project. The sales file, which lacks coordinates, is bridged to parcels through a crosswalk so it inherits the same spatial assignment.

Where the numbers come from

  • Property & parcels — the Polk County Assessor’s AllPolk parcel dataset: assessed values, parcel geometry, neighborhood associations.
  • Sales — county sale records, joined to parcels via the crosswalk.
  • People — the U.S. Census Bureau’s American Community Survey, table B03002 (race and ethnicity as mutually exclusive groups, including Hispanic origin), interpolated to neighborhood boundaries.
  • Amenities — OpenStreetMap, pulled with osmdata: counts and density of cultural, entertainment, retail, and pedestrian-relevant features.
  • Property tax — Polk County Auditor consolidated levy schedules and the Iowa Department of Revenue residential rollback series.

How each score is built

Home value & median sale. Drawn from assessed values and qualifying sale records, summarized per neighborhood.

Walkability, culture, entertainment, retail. Built from the density of the relevant OpenStreetMap features in each neighborhood, then scored on a common scale and spot-validated against known places.

Racial diversity. A Simpson diversity index computed from the B03002 groups, then rescaled linearly to a 0–100 score. This is a deliberate choice: a linear rescale reports Des Moines’s actual level of integration rather than ranking neighborhoods against each other. A percentile rank would have made the most diverse neighborhood look like a 100 even if the city as a whole isn’t very integrated — so it was rejected in favor of honesty.

Economic diversity. The spread of values within a neighborhood, captured with the coefficient of variation, the interquartile range, and the 90th-to-10th-percentile ratio, then expressed as a 0–100 percentile score.

Poverty rate. From the ACS, interpolated to neighborhood boundaries.

Property taxes

Iowa reassesses on a biennial cycle, applies a statewide residential rollback so only part of assessed value is taxable, and then taxes that taxable value at a consolidated levy that varies by taxing district. Ntontan’s tax work transcribes the full district levy matrix from the County Auditor’s schedules and isolates the effect of policy changes across recent assessment years.

Principles & limits

  • Honest over flattering. Scores report real conditions, not a flattering rank for every neighborhood.
  • Some places are left off the map. A handful of boundary slivers have property data but no qualifying sales; they aren’t meaningful as neighborhoods and are excluded rather than shown with thin or misleading numbers.
  • Interpolation has edges. Census figures are estimates interpolated to neighborhood lines and carry the ACS’s own margins of error.
  • Snapshots, not live feeds. Assessments follow Iowa’s reassessment cycle; amenity and census data reflect the vintage they were pulled from.

Specific data vintages (ACS year, assessment and fiscal years) are noted in the project’s processing scripts and should be confirmed against the current build before publishing.