# Understanding Noise Pollution in Boulder — Data Package
**The Quiet Enjoyment Project · September 2026 · Version 1.7**
*(v1.7: the citation-rate figures quoted below now match the report's corrected charge-file counts — 1 charged citation per 46 complaints in South Boulder vs 1 per 320 in the university district, 67 charged party citations citywide; `report_maps_v3.py` gains an optional single-panel share-card render, off by default (`QEP_CARD_OUT`); the hourly-share line notes its denominator; the citywide permanent-resident proxy that `report_stats.py` uses as its outside denominator (77,730, of which 28,010 fall inside the walkshed) is a published constant produced by applying the `age_split.py` rule to every city block — the script that ran that citywide pass is not yet in the package and will be added. No data or derived numbers changed.)*
*(v1.6: the kernel maps now carry full map context — city limits, the hatched university district, labeled arterials and neighborhoods, a scale bar and north arrow — and use the city boundary as their extent, so no call locations fall outside the frame.)*
*(v1.5: `chart_calendar.png` now plots the rest-of-city monthly series beside the walkshed series, so the seasonality contrast is visible on one chart.)*
*(v1.4, after a second adversarial review: `health_exposure.py` now deduplicates
the assessor parcel file by ParcelNo — the city file repeats 2,504 rows — which
lowers the resident-night proximity proxy from 173,441 to 154,048 and the median
nearby-homes figure from 58 (stale, previously hard-coded) to a computed 54;
"school nights" is now reported both as the Sun–Thu weekday shortcut and gated
to actual BVSD student days (~35% / ~54,000); chart annotations are computed,
not hard-coded; `health_charts.py`'s day-name crash is fixed; the stray chart
copies in `data/` are gone; and the walkshed KMZ description no longer embeds a
stale statistic. v1.3: `report_stats.py` computes the 4.1x matched-proxy ratio
directly, with the total-population (2.4x) and superseded mixed-denominator
(4.6x) versions printed as labeled sensitivities.)*

This package contains the datasets, scripts, and figures behind the report
*Understanding Noise Pollution in Boulder* (QEP, September 2026), published per
the report's transparency commitment. Every number derived from these datasets
can be re-derived by running the scripts; the citation and disposition figures
rest on police records published separately (see below). Questions and corrections:
yourneighbor@quietenjoymentproject.org — we publish corrections promptly and
prominently if we err.

`checksums.txt` carries a SHA-256 hash of every other file in the package so
that any copy can be verified against the published original.

## What's in the package

**data/**
| File | What it is | Source |
|---|---|---|
| `noiseb_calls_citywide_2023-01_2026-08.json` | All 9,961 citywide noise complaints (problem type NOISEB), Jan 1 2023 – Aug 14 2026 (pulled the morning of Aug 15; the last record is 3:27 a.m. Aug 15), hundred-block precision | City of Boulder open-data ArcGIS feed, pulled 2026-08-15 |
| `rental_noise_calls_trailing12mo.json` | Noise calls at licensed rental properties (exact address), trailing 365 days | City of Boulder "Rental Property Calls for Service" open-data feed, pulled 2026-08-15 |
| `rental_licenses_parcels_2026-08.json` | 10,830 active licensed-rental parcels with polygon geometry. License-holder name and company fields have been removed — the analysis uses only geometry, address, and license status, and we minimize personal data as a rule; the unabridged layer is the City's own public dataset | City of Boulder rental-license GIS layer, pulled 2026-08 |
| `cu_walkshed_15min.kmz` | The CU walkshed: a 0.75-mile straight-line buffer around the university-district boundary (≈ a 15-minute walk) | QEP, from the City's official subcommunity polygon |
| `streets_osm_arterials.json` | Boulder arterial streets drawn on the maps | © OpenStreetMap contributors, ODbL; Overpass extract 2026-08-31 |
| `walkshed_population_2026-08-31.json` | Walkshed population and group-quarters split | 2020 census PL 94-171, block level (see scripts) |
| `walkshed_age_split_2026-08-31.json` | Walkshed household population by age; the permanent-resident count | 2020 census DHC table P12, block level (see scripts) |
| `term_break_fingerprint_2026-08-31.json` | Complaints per night by academic period, inside/outside the walkshed | Derived; see `term_break_analysis.py` |
| `health_exposure_results.json` | Resident-night proximity estimates (154,048/yr — deduplicated by parcel number and by parcel-night, outlier excluded; 46% on Sun–Thu nights, ~35% gated to actual BVSD student days) | Derived; see `health_exposure.py` |
| `noise_party_calls_trailing12mo.json` | The 3,043 party/noise calls underlying the health exposure figures (2,838 after the standing exclusion of the single-dispute block; every script applies the exclusion) | City of Boulder open data |
| `city_res_parcels.json` | Residential parcel points (parcel number, coordinates, land sqft, assessor neighborhood — no owner information) used for the households-within-earshot join | Boulder County Assessor, pulled 2026-08-22 |
| `citylimits.json` | City of Boulder boundary polygon | City of Boulder GIS |
| `cu_halfmile_buffer.kmz` | Half-mile campus ring (reported for comparison) | QEP |

**scripts/** — the analysis code, as run, with file paths rewritten to the
package layout: run any script from the package root and it reads `./data`
and writes `./data` and `./figures`. The census
scripts (`permanent_pop2.py`, `age_split.py`) expect the public 2020 census
PL 94-171 and DHC bulk files for Colorado unzipped into `./census/` — the
download URLs are in each script's header (~220 MB, not duplicated here) —
and reproduce the published aggregates exactly; the county totals printed by
each script (330,758 residents; 12,094 university group quarters) verify a
correct download.

**figures/** — every chart in the report, at publication resolution.

## Where each headline number comes from

| Claim in the report | Data | Script |
|---|---|---|
| 70% of complaints inside the CU walkshed (6,297 of 9,011) | noiseb + walkshed | `report_stats.py` |
| 10.9 complaint calls/night Thu–Sat in term; one per 44 min; Sept weekends 19; highest night in dataset 38 (Oct 30, 2025) | noiseb + walkshed | `night_stats.py` |
| Term 6.0/night (73%) vs winter break 1.0; move-in fortnight 10.2 (78%) | noiseb + walkshed | `term_break_analysis.py` |
| School-year trend 1,408 → 1,515 → 1,777; spring 2026 +24% | noiseb + walkshed | `report_charts2.py`, `report_maps_v3.py` |
| 3.5x complaints per rental license inside the walkshed (0.26 vs 0.07; 3.3x per unique parcel, 2.8x per dwelling unit) | rental calls + rental parcels + walkshed | `report_charts2.py` |
| 28,010 permanent residents (26% of city population, 36% of its permanent residents), incl. 3,957 children; 4.1x per-permanent-resident rate (matched proxies both sides) | census blocks + walkshed + citylimits | `permanent_pop2.py`, `age_split.py`, `report_stats.py` |
| 55% of complaints 10 p.m.–3 a.m. (of all 9,011 NOISEB records; the report's 57% is the same share within the trailing-12-month party/noise file); hourly distribution | noiseb | `report_charts2.py` |
| ~154,000 resident-nights within 600 ft of a nighttime call (proximity estimate); 46% Sun–Thu, ~35% on BVSD student-day nights; radius sensitivity 43k @300ft / 306k @900ft | party/noise calls + parcels | `health_exposure.py` |
| Maps (noise vs rentals; session vs winter break) | noiseb + rentals + walkshed + OSM streets | `report_maps_v3.py` |

**Citation and disposition figures** (1 charged citation per 46 complaints in South
Boulder vs 1 per 320 in the university district; the barking-dog comparison;
warnings-per-summons) derive from Boulder Police records produced to QEP under
the Colorado Criminal Justice Records Act, three-year window ending October
2025. The BPD charge file is published on our Source Data page
(quietenjoymentproject.org/data.html, under this report; direct path /data/bpd-records/)
in its original form with a CSV copy and a README; the complaint denominators come from
the City's public calls-for-service feed.

## Conventions and notes

- **The excluded block.** One residential block — the 4500 block of 19th
  Street — generated 950 complaints (9.5% of the citywide total) from a single
  ongoing dispute unrelated to the report's subject. It is excluded from every
  figure; the raw data here includes it, and every script applies the
  exclusion in code, so the choice is visible and reversible.
- **Timestamps.** The city's ArcGIS feeds store local wall-clock times in
  epoch fields; the scripts read them accordingly. Calls before 6 a.m. are
  attributed to the prior night.
- **Academic windows.** Term-core Sep 5–Nov 15 and Feb 1–Apr 15
  (ex-spring-break); deep winter break Dec 22–Jan 8. Rates are per night, so
  unequal window lengths compare like-for-like.
- **Privacy and data minimization.** Every record here comes from a source
  the government itself publishes: City of Boulder open-data feeds (which
  carry incident numbers, timestamps, coordinates, and — in the rental-calls
  feed — exact incident addresses, exactly as the City publishes them),
  Boulder County Assessor parcel records, federal census aggregates, and
  OpenStreetMap. The package contains no student records, no University
  production data, and no information about complainants (none exists in
  these sources). Where a source field was not needed for the analysis and
  named individuals — the rental layer's license-holder fields — we removed
  it; the unabridged layer remains available from the City. Incident
  addresses identify properties, not people, and joining these files adds
  nothing beyond what the City's own portal already permits. The
  organizational contact details in this README are published deliberately.

## Licenses and attribution

- City of Boulder open data: per the City's open-data terms.
- U.S. Census Bureau data: public domain.
- Street geometry: © OpenStreetMap contributors, Open Database License (ODbL).
- QEP scripts, derived datasets, and figures: Creative Commons Attribution 4.0
  (CC BY 4.0) — reuse freely with attribution to The Quiet Enjoyment Project.

*The Quiet Enjoyment Project · quietenjoymentproject.org ·
yourneighbor@quietenjoymentproject.org · 885 Arapahoe Avenue, Boulder, CO 80302*