Unexpected Iso-Area polygons far from entry points (QNEAT3 – Iso-Area as Polygons)

[fabiofebocastelli]

Environment

• OS: Windows
• QGIS: 3.44.6-Solothurn
• Python: 3.12.12
• QNEAT3: latest available for QGIS 3.44
• Network CRS: projected (meters, not WGS84)

Problem description

I am using QNEAT3 Iso-Area tools to compute accessibility areas within 300 m walking distance along a street network from multiple entry points (park entrances).

My goal is:

the surface of the city that has at least one entry point reachable within 300 m along the network. Only areas that are reachable within 300 m along the network should be included. No polygon parts should exist where the shortest-path distance to any start point exceeds the specified threshold.

However, I consistently observe areas included in the 300 m Iso-Area polygon that are clearly much farther than 300 m from any entry point, even when validated with shortest-path and distance raster checks.

Workflow used

1. Clean street network:

• Fix geometries
• Reproject to projected CRS (meters)
• Split lines by maximum length (tested 10 m, also tried smaller)

2. Run:

• Iso-Area as Polygons (from Layer)
• Optimization: Shortest
• Size of Iso-Area: 300 m
• Multiple start points (entry points)

3. Post-checks:

• Visually inspect polygons
• Sample the distance raster
• Run Shortest Path from suspect polygon areas to nearest entry point

Observed issues

Some polygon areas:

• Are visibly far from any entry point
• Are confirmed by shortest-path analysis to be >1000–1500 m away
• Still appear inside the 300 m Iso-Area polygon
  In the raster output:
• Some cells far from entry points still show values ≈ 300 m
• Linear artifacts appear outside the expected reachable area
  Splitting the network into shorter segments (10 m) improves results slightly, but:
• The problem does not disappear
• Reducing segment length further would be computationally expensive and still unreliable

See in the screenshot the section of the polygon in the centre that doesn't make sense, as it is far away from any entry point (yellow).

Questions

1. Is this a known limitation of Iso-Area as Polygons in QNEAT3?
2. Is there a recommended workflow to obtain topologically correct accessibility areas for quantitative analysis?