SiteLine: An ATL Metro Area LIHTC Scorer

Preview note — the map library is blocked in this sandbox, so the map falls back to a schematic. Open the file over a local http server to see the real Leaflet map.

The Vortex, Little Five Points

33.7664, −84.3491

Colour Atlanta by

City of Atlanta · 0.05-mile grid (54,424 points).

Projected total · by tract

02040+

Each tract shaded by the score an average site inside it earns. The hard edges are where a point flips across a boundary.

§XIAmenity in each category — hover for detail· click a criterion to refocus

Geocoded via OpenStreetMap (needs internet).

The Vortex, Little Five Points 33.7664, −84.3491

39of 46

Application score

Auto scored

Computed from the site. A few §XI.B deductions can’t be detected from data — abandoned or deteriorated buildings, extensive site mitigation, food-desert — tally those below.

Distances are estimates. The DCA asks that all distances be measured as actual driving or walking distance — along roads and sidewalks, not as the crow flies. Exact route-length computation is a later feature; for now we multiply the straight-line distance by 1.15 as a crude estimate. When a distance sits near a scoring threshold, verify the real route in Google Maps (the DCA’s preferred method, but expensive at scale).

Undetectable undesirable deductions

points to subtract — −2 each (abandoned / deteriorated buildings, mitigation, food-desert)

Developer entry

Credit type

Deal type

Developer-entry criteria

Click a section to open its questionnaire, answer, then Save to add the points — the block turns green and the score above grows. Grey blocks don’t apply to this site or application profile.

Aggregate view

Where the points fall

Every tract in the 11-county region, shaded by the score an average site inside it would earn. Switch the layer above to re-colour the map — read it for the edges, where a parcel across a line gains or loses points.

1,248

tracts scored

33.6

median total

boundary flips

Boundary discontinuities

The most valuable parcels sit on a cliff — where the winning criterion flips:

QCT tract line — basis boost + §XIV ±1

CNI / Purpose Built edge — §XVI ±10

Municipal boundary — §XVIII ±5

Attendance-zone line — §XIII grade swings

Data Sources

Every auto-scored criterion is computed from a specific public dataset, trimmed to the 11-county Atlanta Regional Commission (ARC) footprint and bundled with this app. This page lists what each source is, what it contributes, where to get it — and, deliberately, where it’s weak. Better to show where the bodies are buried than hide them.

Read this first. The web app runs a JavaScript port of the Python scorers on simplified geometry (polygon boundaries thinned by a few tens of metres to keep the download small), so results near a tract/zone/district edge are demo-grade — the Python scorers on full-resolution data remain authoritative. Several sub-items can’t be scored from data at all and are left to the developer. Always verify against the DCA’s own requested sources before relying on a number.

Auto-scored criteria

§XI · Desirable / Undesirable Activities

Nearest-example distance to ~18 amenity categories, minus undesirable uses within 0.25 mi and a food-desert flag. Assembled from many source layers: SNAP retailers (big-box, grocery), Overture Places (retail, restaurants, rec centers, fire/police, worship, post office, nuisance uses), CMS hospitals, NPPES providers/pharmacies, Georgia DECAL childcare, TCSG technical colleges, ParkServe parks, IMLS libraries, FDIC bank branches, EPA/GA-EPD hazards, and the USDA Food Access Atlas.

Sources: USDA SNAP Retailers · Overture Maps · CMS Hospitals · NPPES/NPI · GA DECAL · TCSG · TPL ParkServe · IMLS PLS · FDIC BankFind · GA EPD · USDA Food Access Atlas

Weaknesses: Overture POI completeness and naming vary by category; NPPES addresses can be billing rather than physical locations; the SNAP list is authorization-based (some entries are small stores); ParkServe park classification (≥/<25k ft²) is inferred; the food-desert layer is tract-level and dated (2019 LILA flags). Undesirable categories that need eyes on the ground — abandoned/deteriorated buildings, extensive site mitigation — are not in any dataset and are left to a developer deduction.

§XII · Community Transportation Options

Distance to the nearest transit hub (3+ routes or rail, ≥5 days/week) and nearest stop, built from agency GTFS feeds (MARTA, CobbLinc, Gwinnett County Transit, and others) aggregated via the Mobility Database.

Source: The Mobility Database (GTFS)

Weaknesses: “≥5 days/week” and route counts are derived from each feed’s calendar and trip data — feeds vary in quality and update cadence. Only fixed-route service is captured; on-demand / microtransit is partial. The “hub” definition is a heuristic, and the 6-pt Option-A1 tier (agency-owned land) is a developer/agency documentation item, not scored here.

§XIII · Quality Education Areas

Point-in-polygon over public school attendance zones, scored on CCRPI / Beating-the-Odds qualification and grades covered, by tenancy. Zones and ratings from GreatSchools; CCRPI and Beating-the-Odds from the Georgia Department of Education (via DCA).

Sources: GreatSchools · GaDOE CCRPI

Weaknesses: charter / lottery “zones” are excluded (they’re not as-of-right attendance areas), which is correct per the QAP but means some enrollment options aren’t reflected. Central-Atlanta coverage gaps trigger a nearest-zone fallback (flagged — verify the true attendance zone). New schools may lack CCRPI and use a district substitute. GreatSchools boundaries are best-available, not always the official district map.

§XV · Stable Communities

Five neighborhood indices (environmental health, transit access, jobs proximity, above-poverty share, median income) compared to the pool’s 50th percentile in the site’s census tract. Indicators from the Georgia Housing Needs Assessment (DCA) and Census ACS; tract geometry from Census TIGER.

Sources: Georgia DCA Housing Needs Assessment · Census ACS · Census TIGER 2024

Weaknesses: pool percentile benchmarks are recomputed from the indicator table and depend on its vintage; individual index vintages differ. The “near tract” (≤0.25 mi) borrowing rule is approximated in the web app by distance to tract vertices.

§XVI · Community Designations

Point-in-polygon over HUD Choice Neighborhoods (CNI) grant areas and Purpose Built Communities neighborhoods. In the ARC this is essentially University Homes (Atlanta HA) and Atlanta / East Lake.

Sources: HUD Choice Neighborhoods · Purpose Built Communities

Weaknesses: very few footprints exist in the region and their boundaries are approximate. Being inside makes the 10 pts available only — the award is contingent on the CNI grantee or Purpose Built actually naming the application, which no dataset can tell you.

§XVIII · Previous Projects

LIHTC awards inside the site’s Local Government Boundary (15-yr lookback) and within 1 mile. Award set = HUD National LIHTC Database (≤2019) ⊕ Georgia DCA award / selected-project files (≥2020); municipal boundaries from Census TIGER places.

Sources: HUD National LIHTC Database · Georgia DCA HTC awards · Census TIGER places

Weaknesses: DCA never published 4%/bond award lists for 2020–21 and 2023 — a permanent gap, not a bug. Some older awards have no coordinates and are placed at a seeded random point in-city (valid for the LGB test, excluded from the 1-mile test). Coordinate precision varies (exact / geocoded / HUD-matched / OSM / random). If you suspect a nearby award, verify directly with DCA.

§XIX · Housing Needs Characteristics

Other-Metro-pool only: not-QCT + ≥45% severe-housing renters + growing county, with a §XV/§XIV bonus. Severe housing problems from HUD CHAS; growth from Census ACS (population) and BLS (employment); QCT from HUD.

Sources: HUD CHAS · U.S. BLS · HUD Qualified Census Tracts

Weaknesses: never applies to the Atlanta-Metro pool (correct per QAP). CHAS is a multi-year ACS special tabulation and lags; the county-growth thresholds are sensitive to the ACS/BLS vintage used.

§XXII · Historic Preservation

Point-in-polygon over National Register of Historic Places district / resource boundaries (125 in the ARC).

Source: NPS National Register of Historic Places

Weaknesses: NR boundary geometry completeness varies. As with §XVI, being inside makes the 2 pts available, not guaranteed — they require historic tax credits and adaptive reuse, a developer commitment.

Cross-cutting caveats

Simplified geometry. Tract, zone, place, and district polygons are thinned for the web bundle; a point within a few tens of metres of a boundary may score differently than the full-resolution Python scorer.
Developer-entry criteria (§III–§XXV, non-auto) are self-reported. The tool structures the questions and computes points from your answers — it does not verify them.
§XX Economic Development, §XIV Revitalization documentation, and other narrative items depend on letters and adopted plans no dataset captures; they’re guided-input, not auto-scored.
QAP interpretation. Our reading of the 2024–2025 QAP may not align exactly with the DCA’s (or InvestATL’s). Treat scores as guidance and check every one with a qualified professional.