Property Values Along the Northaven Trail

Methodology & Data Sources

We analyzed public appraisal data from the Dallas Central Appraisal District (DCAD) and parcel boundary data from the DCAD GIS portal. Each parcel was assigned to a distance band based on the distance from its centroid to the Northaven Trail centerline using the NAD83 / Texas North Central (EPSG:2276) projected coordinate system.

Parcels are also grouped into Trail segments using major road boundaries: US 75, Hillcrest, Preston, Inwood, Midway, Marsh, Royal, Webb Chapel, and Monroe. Each segment is further split into three distance rings — within ½ mile of the Trail, ½–1 mile, and 1–2 miles — drawn on the map with a solid, dashed, and dotted boundary respectively. The unsplit end zones (East of US 75 and West of Monroe) wrap the Trail's two endpoints; note that West of Monroe has no Trail frontage — its homes are near the western trailhead.

Values shown are appraised market values from the certified appraisal roll for each tax year. Median values are used for most comparisons because they are less affected by very high-value outlier properties.

Before/after opening comparison: Phase 2 opened in stages — Preston Rd to Marsh Ln in May 2019, and the section west of Marsh Ln in 2020 — and DCAD values are certified as of January 1, so the rolls preceding each opening capture those neighborhoods before their stretch of Trail existed. The "Before the Trail Opened" section compares single-family value growth in the Trail area (within ½ mile) against the 1–2 mile control ring over those pre-opening and post-opening windows, using only matched parcels — parcels with a positive, non-exempt value in both endpoint years — so changes in the parcel mix don't distort the comparison. Growth rates are annualized for windows of different lengths.

Robustness of the headline comparison: The lead metric compares growth in the median single-family value across all parcels on each year's roll, so in principle it could be moved by changes in the parcel mix — new construction, teardowns, replats — rather than by appreciation of existing homes, or by the choice of starting year. Four checks, recomputed automatically from parcel-level records for the full — window, show that it is not. (1) Constant-parcel basis: restricting the comparison to matched parcels (— Trail-area and — control-ring homes on the roll in both endpoint years) gives a gap of —. (2) Excluding teardown-rebuilds: additionally removing the — matched parcels that show the appraisal signature of a demolition and rebuild — a structure worth at least $50,000 in the base year whose improvement value later collapsed below $20,000 (or below 20% of its base level) and ended at more than double its base level — gives —. (3) Land value only: comparing the median of each parcel's own land-value growth — land is unaffected by construction or remodeling, and the per-parcel form is used because pooled land medians are sensitive to the mix of housing types; endpoint years only, because DCAD reprices land tables in periodic steps rather than annually — gives —. (4) Baseline sensitivity: the two cohorts reached their early-2010s trough at slightly different times, so the anchor year matters in principle; re-anchoring the headline comparison at — — each measured over all parcels on that year's roll, exactly as the headline is — moves the gap by a few points without changing the story. The first two checks use the same growth-of-the-median measure as the headline; the land check is a different but composition-proof measure; and all of the checks point the same way as the headline gap for the same window shown in the overview.

Dallas City Council districts are a secondary reporting geography shown in the district section (they are not drawn on the map). Each parcel is assigned to its current council district by a point-in-polygon spatial join between the parcel centroid and the official City of Dallas council-district boundaries, reprojected to EPSG:2276. A parcel whose centroid falls outside every district is left unassigned rather than forced into the nearest one. District figures are recalculated directly from parcel-level records: the headline is aggregate value growth (the change in the combined value of all valid parcels), reported alongside the median parcel growth. The Dallas comparison is expressed in percentage points. Assignments are tied to a boundary version, so districts can be recomputed after redistricting without rebuilding historical valuation data. See the project documentation for the full method.

Added city & county tax revenue (estimates): we compare the Trail Area's total appraised value growth since the baseline year against a counterfactual in which it had only grown at the same rate as the 1–2 mile control cohort. The excess value is multiplied by the adopted property tax rate for the viewed year — the City of Dallas rate for the city metric and the Dallas County rate for the county metric (Dallas County is a major funder of the Trail). Each metric uses only that entity's rate; DISD, Parkland Hospital, and Dallas College taxes are excluded. Rates are verified against the Dallas County Tax Office's and Texas Comptroller's published adopted rates. The cumulative card is opening-aware: each trail section only counts from the year after its stretch opened (US 75–Preston from the 2012 roll, Preston–Marsh from 2019, west of Marsh from 2020), with each section's excess measured against its own 1–2 mile ring and taxed at that year's adopted city + county rates — growth from before a section existed is never attributed to the Trail. Because the calculations use appraised market value rather than taxable value (which is reduced by homestead exemptions and appraisal caps), these should be read as upper-bound estimates.

Important: This analysis is descriptive. It shows how property values near the Trail have changed over time compared with broader Dallas averages. It does not prove that the Trail caused those changes. Many factors influence property values.