Punggol Digital District
Explore the reviewed indicators for the hand-drawn district and its H3 cells. District results use the PDD boundary, not neighbouring subzone values; accessibility and ecosystem-connectivity summaries use overlap-area-weighted means. Preliminary, modelled and pending results remain clearly identified.
1. District snapshot
Explore the review outputs and colleague's supplied final distances, not a final canonical release. The linked maps clip R10 cells to the hand-drawn PDD boundary, shown as an orange outline. Choose indicators or category scores to compare cells within PDD. Filters update the distributions, scatter, maps and matrix together. Category scores use fixed PDD quartiles; shorter access distances receive higher points.
2. Indicator presentation
3. Linked cell explorer
Methods and data
Detailed source notes and earlier analysis charts
The colleague's five accessibility layers are merged by cell ID and shown using the supplied FinalDist_m values only. Every PDD cell matched. These final distances include network results and the colleague's Euclidean proxies for network gaps; we do not recalculate or fill them. Histograms exclude missing final values. Source year, routing settings and destination definitions remain to be documented; the import does not replace the approved canonical methods.
Relationships between the presentation indicators
A single district is one observation at district support, so nothing can be correlated there. The R10 cells inside the boundary are the only population available, and only indicators that genuinely vary across those cells may enter. Outcome indicators are shown in bold throughout this section.
Capacity against outcomes
The matrices give a coefficient; this gives the shape behind it, which is where a curve, a cluster or a single influential cell shows up. It follows the same scatter used on the Subzone explorer, so the two pages read the same way. The default pairs a capacity indicator against an outcome.
Cells by quartile
Quartiles are cut across the cells inside this boundary, on the same PDD-relative basis as the histogram cut points on the maps above — Q4 is the best band, so an accessibility distance is banded with short distances at the top. These are display bands, not canonical framework scores, and they say nothing about where a cell would sit citywide.
Cells strong on several indicators at once
This is an intersection, not a shading: a cell qualifies only if it sits in the top quartile on every indicator selected. That answers where the district is strong on all of them together, which is not something the separate maps above can show.
City context and method notes
Ecosystem connectivity’s PDD summary is the mean of raw nearest-patch normalized betweenness, weighted by each R10 cell’s area inside the drawn boundary. Valid zeros count; missing values are excluded and area coverage is reported. The full Singapore habitat graph is unchanged: this is not centrality recalculated on a PDD-only graph, nor evidence of species movement. The adapted vegetated-habitat proxy includes managed vegetation.
Land-use compactness is a subzone/district indicator: one Global Moran's I calculated directly within the boundary. R10 footprint coverage is the primary input and native R11 is sensitivity. H3 coverage and Local Moran/LISA are diagnostic layers, not H3 versions of the district outcome. The reviewed building source is 2023.
Green Amenity Access here is a preliminary walking-distance diagnostic. The framework outcome we intend to calculate is the year-on-year change in the share of residents within 400 m walking distance of greenery. Counting qualifying H3 centroids does not measure that resident share. These maps use a 2026 network with 2024 vegetation; historical access requires an explicit network assumption.
Effective mesh size is a Singapore-wide statistic, not a local PDD value. It uses all natural components, strict <100 m grouping, a total-natural-area denominator, and no added road barriers. It is not filled into the H3 or PDD maps.
District GPR uses whole-boundary support, including water, at 3.62 m floor height as its approved headline. The four paired GFA/area variants below retain non-water and 3.2 m sensitivities. Universal Greenspace is now a modelled PDD presentation candidate using the new per-image-median NDVI source. Its population allocation is sensitive to the coarse WorldPop prior; it is not a canonical replacement or an observed PDD population. Access to Nature now includes the colleague's supplied H3 final distances; source year, routing settings and destination eligibility still require documentation.
Earlier framework explorer — legacy snapshot, not the selected presentation outputs
Punggol Digital District
Everything on this page is calculated for the boundary drawn for this study — not for an official subzone, and not by borrowing a neighbour's value. Where a figure is compared against Singapore, the district is placed among the official subzones without ever joining them.
1. The district's values
Each row shows how the figure was produced. The three routes do not carry equal evidential weight, so the label is part of the reading.
2. Inside the boundary
3. Relationships between indicators inside the district
A single district is one observation at subzone support, so nothing can be correlated there. The cells inside the boundary are the only population available, and only indicators that genuinely exist at H3 may enter.