Layer guide
What every layer on this map is, and where its data comes from. The left panel is ordered by the seven-step TRH 26 process in the workshop pack: 1 fix the network → 2 split into links → 3 rural/urban → 4 three criteria → 5 class number → 6 test connections → 7 sense-check & data requests. Steps 1–2 and 7 are desk work; the room does 3–6.
Local municipality boundaries (Step 2). The 19 Free State local and metro municipalities from the Municipal Demarcation Board 2026 boundary service, each carrying its Census 2022 profile (population 2011→2022 and growth, households, dwellings, water, electricity, education), Microsoft building-footprint count, SARS 2024 assessed-taxpayer figures and MIIF category. Hover for the headline, click for the full sheet. Source: FS_CHRIS data pack (5 Sep 2026) — Stats SA, MDB, SARS.
Built-up footprint (Step 3). Where the province is actually subdivided and built on: 1 231 596 machine-inferred building footprints (Microsoft Global ML Building Footprints, release 2026-07-24, CDLA-Permissive-2.0) binned to 0.5 km cells; a cell is shown when it holds 12 or more footprints (≈50 per km²). This is the on-screen test for TRH 26 §4.3 — inside the footprint a link is urban (U classes) unless it is a through-way whose function stays rural. Footprints are ML output, not a cadastre: use them to locate the urban edge, not to measure it.
Settlement size tiers (Step 4). Each town's built-up cluster (contiguous footprint cells) with its population from the Census-2022-calibrated 1 km population grid (WorldPop base, control-totalled to Stats SA municipal counts). Ring colour and size give the trip-generator rung of the decision ladder: blue >500 000 or metro seat (Class 1 generator), red >25 000 (Class 2), orange 2 000–25 000 (Class 3), green <2 000 (rural settlement, Class 4 destination). Bloemfontein ≈480 000 in-cluster (metro seat → tier 1), Phuthaditjhaba ≈306 000, Welkom ≈202 000, Botshabelo ≈177 000, Sasolburg ≈109 000, Kroonstad ≈96 000. Closes gap G5 (town-level population tiers) of the workshop's Annexure B.
Schools (Step 4). 1 450 Free State schools from the National Governance System / CSIR facilities layer (DBE EMIS), joined on EMIS number to the DBE address list of 2 March 2017 for learner numbers and quintile (1 229 matched). Dot size follows school type: primary smallest, secondary and technical largest. Hover for the headline, click for the sheet. Why it is here: TRH 26 reads a road's class from what it connects, and a school is the commonest destination at the end of a Class 4 collector — this layer is the evidence for the collector / local decision that the register makes worst. Supplied by Christopher du Plessis, 9 Sep 2026.
Health facilities (Step 4). 317 clinics, health centres and hospitals inside the province: 276 from the WHO-KEMRI 2019 public-facility register (tiered — 245 Tier 2 clinics and health centres, 30 Tier 3 provincial hospitals, 1 Tier 4 central hospital) and 41 from healthsites.io / OpenStreetMap (private hospitals, pharmacies, untiered clinics). Dropped: 44 OSM points flagged as probable duplicates of a WHO record, and 135 points that fall in neighbouring provinces. Size follows tier. A Tier 3 / 4 hospital is a regional generator (Class 2–3 reach); a clinic is a local one (Class 4 / 5). Supplied by Christopher du Plessis, 9 Sep 2026.
SANRAL — owned centrelines. Real road geometry from the SANRAL RAMS Route_Section shapefile, blue where the OWNER field says NRA. Line width follows the recorded cross-section (freeway / dual / single). Click a section for chainage, lanes, pavement and its indicative TRH 26 class.
Province — owned centrelines. The same shapefile where OWNER = PROV, green. Mixed-ownership routes (R26, R34, R59, R64…) reflect the 2024/25 transfers to SANRAL.
Provincial paved — RRAMS. 548 links, 5 379 km from the district RRAMS shapefiles (Jan 2026), P/S/A numbering, each with a RISFSA class (R1–R5) and from/to description.
Provincial gravel — RRAMS. 1 355 links, 20 583 km, the S-route gravel network — the bulk of the province's roads, classified per link by SMEC.
Colour by Ownership / Surface / Class. Ownership shows custodian (blue/green). Surface recolours by pavement type on the RCAM convention: surfaced (blacktop and concrete) black, unsurfaced Light Sienna. Class recolours EVERY road layer by its RISFSA / TRH 26 class on the RCAM palette — SANRAL trunk, RRAMS paved and gravel, tertiary, L-roads, municipal and metro streets, and the traffic-count overlays (which keep ADT width but drop their truck-share colours): Class 1 Cretean Blue, 2 Poinsettia Red, 3 Electron Gold, 4 Leaf Green, 5 Medium Coral, 6 40% Grey, undefined 40% Grey dashed. Line weight follows the class too (2 : 2 : 2 : 1 : 0.5 : 0.4), so in Class mode the hierarchy is visible at a glance and the thin-collector problem shows itself: very little green between the arterials and the coral local web.
RCAM symbology source. NW_RCAM Road Classification Symbology, Royal HaskoningDHV for the North West Department of Public Works and Roads, 25 March 2020 — the display convention for RCAM classes, surface types and urban boundaries, using the Esri ArcMap standard colour set (values from Esri's published RGB table: Cretean Blue 0,130,255 · Poinsettia Red 230,0,0 · Electron Gold 255,170,0 · Leaf Green 56,168,0 · Medium Coral 245,122,122 · 40% Grey 156,156,156 · Light Sienna 205,170,102). Two parts of the convention are not drawn here: the rural / urban line style (rural classes dashed 1-long-1-short, urban solid, U4a/U5a/U6b nautical-dashed) — 99.8% of this register is rural, so dashing it would add nothing but ~115 000 dashed polylines to redraw; and the urban-boundary polygons (Residential Olivenite Green 168,168,0 · Residential commonage Light Apple 163,255,115 · Commercial Anemone Violet 169,0,230 · Commercial commonage Fuschia Pink 255,115,223, outline 1) — the province has no RCAM land-use boundary layer yet; the built-up footprint above is the interim urban test. Adopt these when the boundaries are produced.
Measured traffic (SANRAL AADT). 254 count segments on real centrelines. Width = AADT, colour = truck share (green <10% to dark red 40%+). Counts 1995–2023 — pre-2018 segments carry a warning.
Project ADT counts 2026 (TMH-18). This project's own counts (Meridian/V N-7979/8007/8056, validated clean; Feb–Jul 2026), painted onto the RRAMS road geometry exactly like the SANRAL traffic layer: width = ADT, colour = heavy-vehicle share. 385 roads painted (~7 000 km). Small hollow rings = stations counted with no data recorded; dots = the few roads not in the mapped layers; dashed = QA flag. Same width/colour scale as the SANRAL layer, so the two are directly comparable.
All-FS networks (owner-aligned). Four coverage layers from the department's All FS data folder, each attributed to its owner and coloured accordingly in Ownership mode: Provincial tertiary T-roads (green thin, 17 852 km — the network that reconciles the provincial denominator), District-owned L-roads (cyan — auth LPDM, the first identified district roads), Municipal streets (purple thin — Mangaung + Fezile Dabi georams + TMDM + Xhariep + Lejweleputswa FSDC18NETWORK Dec 2025, ~12 836 km with RCAM classes), and Unproclaimed candidates (olive dashed, off by default — 9 773 km of OSM-derived roads with no registered owner). All respond to Class colour mode and the class filter chips; hover or click a network for its summary.
TRH 26 principle check (current classes). Automated audit of the recorded classes against the manual's structural rules, independent of traffic: red ⊗ = forbidden junction connections (56 — mostly R5 access roads meeting R1/R2 arterials directly, s4.6.2), amber ⊗ = exceptional Class 1–3 connections to verify (12), and ink-dotted roads = class inconsistent with its own length (535 — led by 441 R5 'local access' roads longer than 15 km, the longest 134 km). Province-wide proportions vs the manual's benchmarks: R1+R2 at 7.2% against 2–4%, R1–R3 at 17.7% against 6–12%, R4 at 15.7% against 20–25% — the network is top-heavy on arterial labels and thin on collectors.
Suggested RISFSA changes (TRH 26 review). Purple dashed = upgrade candidates (27), teal dashed = downgrade candidates (17), grey dashed = the 13 roads where the two class sources disagree with no volume trigger. Flagged where the 2026 measured ADT falls outside the TRH 26 typical band for the road's recorded class (s4.4), or where the SMEC shapefile and RAMS export disagree. Suggestions are review candidates only — TRH 26 classifies on function, never volume; the panel of each road says exactly why it was flagged.
Strategic traffic generators & major developments. 22 areas that attract trips (rings — dashed red = not documented in any municipal plan) plus 5 major developments (magenta diamonds — dashed outline = position approximate): Mabele Fuels, Virginia Gas Project, NAMPO Park, Parys and Clarens tourism.
Demographics — Census 2022. Population per municipality, plotted at the seat: more people, bigger circle. Mangaung 811 431 down to Tokologo 29 455.
Strategic development density. How many strategic developments each municipality's SDF and IDP record (270 total across 19 municipalities).
Municipal network size. Published municipal road length per LM, where a figure exists. Click for condition quotes.
District municipality boundaries. Municipal Demarcation Board polygons: DC16, DC18, DC19, DC20 and Mangaung. Click one for its road-planning picture.
Magisterial districts. 53 town catchments from the SANRAL_Districts shapefile — useful as service-area proxies.
Towns (85) and Townships & settlements (90). Towns are solid dots with always-on labels; townships are hollow rings, plotted adjacent to their parent town (position approximate), labels appear as you zoom in.
Toll plazas, hubs, border posts. Toll plazas (red), logistics/economic hubs (purple), border posts (green), tourism nodes (orange). Tolls matter: Verkeerdevlei and Vaal drive the R721/R82/Brandfort diversion routes.
Borrow pits & quarries. 164 SANRAL materials sites with RDS numbers — materials planning for any upgrade the classification triggers.
Road declarations. 162 gazette-numbered declaration polygons (44 of 2024, 34 of 2025) — the legal record that settles custodianship.
Filter by TRH 26 class. The chips filter the trunk network by indicative class and the RRAMS links by their RISFSA class (R# read as Class #).
Base map — Streets, Topo, Satellite. Live tiles (internet required), switched with the chips under View: Streets is OpenStreetMap, Topo is Esri's World Topographic Map (relief, contours, place names), Satellite is Esri World Imagery (Maxar / Earthstar). Your choice is remembered in this browser. All three serve to zoom 19, about 30 cm per pixel over the province — zoom right in to check a junction, a bypass or the edge of a settlement against the classification. When a base map is on, district fills go translucent so it reads through. Attribution is shown in the map corner.