34:["$","title",null,{"children":"Day 9: Urban–rural | Xavier Olive | Observable"}] 3c:T533,## How it works Counting doctors per commune would mislead: two communes in three have no general practitioner at all, and most of them are no desert, their inhabitants simply drive to the next village. The APL, computed by the DREES, the statistics office of the ministry of health, answers this. It counts every practitioner within a short drive (up to 20 minutes for GPs, with a weight that falls with the distance), weighs them by their actual activity in the year (a part-time doctor counts less than a full-time one), and divides by a population standardised by age, since older people see their doctor more often. For general practitioners, the APL is a number of consultations a year per inhabitant, 3.3 on average in France in 2024, and a commune below 2.5 is commonly counted as under-served. For nurses, physiotherapists, dentists and midwives, the APL is a number of full-time practitioners per 100,000 inhabitants (per 100,000 women for midwives), and there is no official threshold: the map takes two thirds of the national average, which for GPs gives 2.5 again. The colour scale is logarithmic and centred on the threshold: brown below, blue-green above. Paris, Lyon and Marseille are given by arrondissement in the APL; they are merged here into one commune each, weighted by population. Mayotte is not covered.3d:T9b8,## How these were recomputed Each practice shares its practitioners among the population within reach; each commune then adds up the shares of the practices within reach, with a weight that falls with the distance in both steps. Reach is 45 km for the two specialists, seen a few times a year at most, and 30 km for speech therapists, whose patients, mostly children, often come every week. This is a two-step floating catchment, the method behind the APL, but cruder: it counts heads rather than actual activity, and straight lines between communes rather than driving times, hence the round halos around the towns. The directory (extraction of 6 November 2026) lists 3,525 dermatologists, 5,895 ophthalmologists and 29,967 speech therapists with a place of practice, a practitioner working in several places counting as a fraction in each. These practitioners gather in the towns, which serve whole regions, so the threshold is half the national average: 2.6 dermatologists, 4.3 ophthalmologists or 22 speech therapists within reach for 100,000 inhabitants. **Is it as bad as it looks?** Distance alone is reassuring. In mainland France, 73 % of the population lives within 10 km of a dermatologist, in a straight line, and only 2.4 % more than 30 km away; 81 % and 0.5 % for ophthalmologists; 98 % live within 10 km of a speech therapist. The shortage is in numbers rather than in distance, which is why patients wait for months: about 5 dermatologists for 100,000 inhabitants with a place of practice in the directory, and fewer still, about 3.3, in published counts of those actually practising. Below the threshold live 28 % of the population for dermatologists, 14 % for ophthalmologists and 15 % for speech therapists, from 5 %, 1 % and 4 % in the large cities to 77 %, 51 % and 55 % in the most sparsely settled countryside. The directory lists no dermatologist at all in the Nièvre, as its members of parliament have pointed out, and fewer than 1 for 100,000 inhabitants in the Ariège, the Eure and the Meuse, against 26 in Paris. Ophthalmologists are better spread, but the Haute-Saône, the Haute-Marne and the Meuse have 2 to 3 for 100,000 inhabitants, against 34 in Paris. Speech therapists tell the most urban story of the three: the Seine-Saint-Denis, next to Paris, has fewer of them per inhabitant (20 for 100,000) than the Cantal, close to the rural Indre, Deux-Sèvres and Allier (17 to 19), against more than 70 in the Hérault, the Rhône or the Indre-et-Loire.3e:T2afb,// ── Helpers: colour, legend, map, urban–rural shares ─────────────────────────── // The map is drawn on a canvas: about 35,000 communes, projected once per width (prepare), coloured for each // profession (drawMap); hovering a commune shows its figures. apl = { const PROFESSIONS = { gp: {label: "General practitioners", unit: "consultations a year per inhabitant", digits: 1}, nurse: {label: "Nurses", unit: "full-time nurses per 100,000 inhabitants", digits: 0}, physio: {label: "Physiotherapists", unit: "full-time physiotherapists per 100,000 inhabitants", digits: 0}, dentist: {label: "Dentists", unit: "full-time dentists per 100,000 inhabitants", digits: 0}, midwife: {label: "Midwives", unit: "full-time midwives per 100,000 women", digits: 0}, // recomputed here from the RPPS directory (not published by the DREES), shown on the second map derm: {label: "Dermatologists", unit: "dermatologists within reach per 100,000 inhabitants", digits: 1, recomputed: true, note: "Not published by the DREES: recomputed here from the RPPS directory, see below"}, ophtha: {label: "Ophthalmologists", unit: "ophthalmologists within reach per 100,000 inhabitants", digits: 1, recomputed: true, note: "Not published by the DREES: recomputed here from the RPPS directory, see below"}, speech: {label: "Speech therapists", unit: "speech therapists within reach per 100,000 inhabitants", digits: 0, recomputed: true, note: "Not published by the DREES: recomputed here from the RPPS directory, see below"}, }; /** d3 wants exterior rings clockwise: a polygon whose area exceeds a hemisphere is wound the other way */ function rewind(d3, feature) { const g = feature.geometry; if (!g) return feature; const polygons = g.type === "Polygon" ? [g.coordinates] : g.coordinates; for (const rings of polygons) { if (d3.geoArea({type: "Polygon", coordinates: rings}) > 2 * Math.PI) for (const ring of rings) ring.reverse(); } return feature; } const DENSITY = ["Large cities", "Mid-sized towns", "Suburbs", "Small towns", "Rural towns", "Dispersed rural", "Very dispersed rural"]; /** National mean (weighted by population) and the threshold of under-supply: * 2.5 consultations for GPs, as used by DREES and the ARS; two thirds of the national mean for the other * primary care professions (for GPs, two thirds of the mean is 2.5 as well); half the national mean for the * layers recomputed from the RPPS directory, whose practitioners gather in the towns that serve whole regions. */ function stats(communes, key) { let s = 0, w = 0; for (const c of communes) { const v = c.properties[key]; if (v != null && c.properties.pop) { s += v * c.properties.pop; w += c.properties.pop; } } const mean = s / w; return {mean, threshold: key === "gp" ? 2.5 : PROFESSIONS[key].recomputed ? mean / 2 : (2 / 3) * mean}; } /** Colour: brown below the threshold (the "desert"), blue-green above, on a log scale centred on the threshold */ function colour(d3, threshold) { const scale = d3.scaleDiverging([-1.6, 0, 1.6], (t) => d3.interpolateBrBG(t)); return (v) => (v == null ? "#d9d4cc" : v <= 0 ? "#543005" : scale(Math.max(-1.6, Math.min(1.6, Math.log2(v / threshold))))); } /** Projection and paths, computed once per width: metropolitan France in a conformal conic, * the four overseas départements in boxes on the left; a Delaunay triangulation of the centroids finds the commune under the pointer */ function prepare(d3, data, width) { const communes = data.communes, deps = data.departements; const metro = communes.filter((c) => !c.properties.dep.startsWith("97")); const left = width * 0.17, collection = {type: "FeatureCollection", features: metro}; const main = d3.geoConicConformal().parallels([44, 49]).rotate([-3, 0]).fitWidth(width - left - 10, collection); const [[, y0], [, y1]] = d3.geoPath(main).bounds(collection); const height = Math.ceil(y1 - y0) + 20; main.translate([main.translate()[0] + left, main.translate()[1] - y0 + 10]); const box = Math.min(width * 0.13, (height - 20) / 4 - 6), gap = (height - 20 - 4 * box) / 3; const overseas = [["971", "Guadeloupe"], ["972", "Martinique"], ["973", "Guyane"], ["974", "La Réunion"]].map(([dep, name], i) => { const features = communes.filter((c) => c.properties.dep === dep); const x0 = 10, yy = 10 + i * (box + gap); const projection = d3.geoMercator().fitExtent([[x0 + 4, yy + 14], [x0 + box - 4, yy + box - 4]], {type: "FeatureCollection", features}); return {dep, name, features, projection, x0, y0: yy, box}; }); const views = [{features: metro, projection: main, deps: deps.filter((d) => !d.properties.code.startsWith("97"))}, ...overseas.map((o) => ({...o, deps: deps.filter((d) => d.properties.code === o.dep)}))]; const items = [], borders = new Path2D(); for (const view of views) { const path = d3.geoPath(view.projection); for (const c of view.features) items.push({feature: c, path: new Path2D(path(c) ?? ""), centroid: path.centroid(c)}); for (const d of view.deps) borders.addPath(new Path2D(path(d) ?? "")); } // picking: the commune whose centroid is nearest, or one of its neighbours, that contains the point const delaunay = d3.Delaunay.from(items, (d) => d.centroid[0] || 0, (d) => d.centroid[1] || 0); const test = document.createElement("canvas").getContext("2d"); const find = (x, y) => { const i = delaunay.find(x, y), seen = new Set([i]), queue = [i]; for (let k = 0; k < queue.length && k < 40; k++) { const j = queue[k]; if (test.isPointInPath(items[j].path, x, y)) return items[j]; for (const n of delaunay.neighbors(j)) if (!seen.has(n)) { seen.add(n); queue.push(n); } } return null; }; return {width, height, items, borders, overseas, find}; } function drawMap(prepared, key, {threshold}, d3) { const {width, height, items, borders, overseas, find} = prepared, dpr = devicePixelRatio || 1; const canvas = document.createElement("canvas"); canvas.width = width * dpr; canvas.height = height * dpr; canvas.style.width = width + "px"; canvas.style.maxWidth = "100%"; const ctx = canvas.getContext("2d"); ctx.scale(dpr, dpr); const fill = colour(d3, threshold); ctx.lineWidth = 0.3; for (const {feature, path} of items) { ctx.fillStyle = ctx.strokeStyle = fill(feature.properties[key]); ctx.fill(path); ctx.stroke(path); // the stroke closes the hairline gaps between communes } ctx.lineWidth = 0.6; ctx.strokeStyle = "rgba(255,255,255,0.8)"; ctx.stroke(borders); ctx.font = `${Math.max(9, width * 0.012)}px system-ui, sans-serif`; ctx.fillStyle = "#666"; ctx.strokeStyle = "#ccc"; ctx.lineWidth = 0.5; for (const o of overseas) { ctx.strokeRect(o.x0, o.y0, o.box, o.box); ctx.fillText(o.name, o.x0 + 4, o.y0 + 11); } // hover: the commune, its value and its class const tip = document.createElement("div"); tip.style.cssText = "position:absolute;pointer-events:none;background:white;border:1px solid #ccc;border-radius:3px;padding:4px 6px;font:12px system-ui,sans-serif;display:none;white-space:nowrap;box-shadow:0 1px 3px rgba(0,0,0,.15)"; const wrap = document.createElement("div"); wrap.style.position = "relative"; wrap.append(canvas, tip); const fmt = (v) => (v == null ? "no data" : v.toFixed(PROFESSIONS[key].digits)); canvas.addEventListener("pointermove", (e) => { const r = canvas.getBoundingClientRect(), x = ((e.clientX - r.left) / r.width) * width, y = ((e.clientY - r.top) / r.height) * height; const item = find(x, y); if (!item) { tip.style.display = "none"; return; } const p = item.feature.properties, v = p[key]; tip.innerHTML = `${p.name} (${p.dep})
${fmt(v)} ${PROFESSIONS[key].unit}
` + `${p.pop?.toLocaleString("en") ?? "?"} inhabitants · ${DENSITY[p.dens - 1] ?? "?"}` + (v != null && v <= threshold ? `
below the threshold` : ""); tip.style.display = "block"; const left = e.clientX - r.left + 12, top = e.clientY - r.top + 12; tip.style.left = Math.max(0, Math.min(left, r.width - tip.offsetWidth - 4)) + "px"; tip.style.top = top + "px"; }); canvas.addEventListener("pointerleave", () => (tip.style.display = "none")); return wrap; } /** Legend: the colour ramp on a log scale, with the threshold marked */ function legend(d3, key, {threshold}, width = 360) { const fill = colour(d3, threshold), h = 58, m = 12; const x = d3.scaleLog([threshold / 3.03, threshold * 3.03], [m, width - m]).clamp(true); const svg = d3.create("svg").attr("width", width).attr("height", h).attr("viewBox", [0, 0, width, h]).style("font", "11px system-ui, sans-serif").style("overflow", "visible"); const n = 120, values = d3.range(n).map((i) => x.invert(m + ((width - 2 * m) * (i + 0.5)) / n)); svg.append("g").selectAll("rect").data(values).join("rect").attr("x", (v) => x(v) - (width - 2 * m) / n / 2).attr("y", 24).attr("width", (width - 2 * m) / n + 0.5).attr("height", 10).attr("fill", fill); const ticks = key === "gp" ? [1, 2, 2.5, 3, 4, 5, 7] : x.ticks(6).filter((t) => [1, 2, 5].includes(+t.toPrecision(1)[0])); svg.append("g").selectAll("text").data(ticks).join("text").attr("x", x).attr("y", 46).attr("text-anchor", "middle").attr("fill", "#555").text((t) => (key === "gp" ? t : d3.format(",")(t))); svg.append("line").attr("x1", x(threshold)).attr("x2", x(threshold)).attr("y1", 20).attr("y2", 36).attr("stroke", "#222").attr("stroke-width", 1.2); svg.append("text").attr("x", x(threshold)).attr("y", 16).attr("text-anchor", "middle").attr("fill", "#222").text(`threshold ${key === "gp" ? "2.5" : threshold.toFixed(PROFESSIONS[key].digits)}`); svg.append("text").attr("x", width / 2).attr("y", h).attr("text-anchor", "middle").attr("fill", "#777").text(PROFESSIONS[key].unit); if (PROFESSIONS[key].note) svg.attr("height", h + 16).append("text").attr("x", m).attr("y", h + 15).attr("text-anchor", "start").attr("fill", "#8c510a").style("font-weight", "bold").text(PROFESSIONS[key].note); return svg.node(); } /** Share of the population below the threshold, by class of the density grid */ function byDensity(communes, key, threshold) { const rows = DENSITY.map((label, i) => ({label, cls: i + 1, pop: 0, low: 0})); for (const c of communes) { const p = c.properties, v = p[key]; if (v == null || !p.pop || !p.dens) continue; rows[p.dens - 1].pop += p.pop; if (v <= threshold) rows[p.dens - 1].low += p.pop; } return rows.map((r) => ({...r, share: r.low / r.pop})); } return {PROFESSIONS, rewind, DENSITY, stats, colour, prepare, drawMap, legend, byDensity}; }3f:[{"id":"1ae0584203da0f85c02b2bf6cc56a96ba286acbc698bc88dc9dcd8139cdd72563757ec314ed6ce3f68aaac7b310c951d9eed0e89c91bdc87357946863932f6ab","href":"https://static.observableusercontent.com/files/1ae0584203da0f85c02b2bf6cc56a96ba286acbc698bc88dc9dcd8139cdd72563757ec314ed6ce3f68aaac7b310c951d9eed0e89c91bdc87357946863932f6ab","name":"france-health-v3.json.gz","type":"application/octet-stream","lastModified":1794061151256,"size":2634175}] 38:{"id":"8de237a689f36b8b","slug":"30-day-map-challenge-2026-day-9-urban-rural","title":"Day 9: Urban–rural","thumbnail":"304f0dd6835741b741d003e35fd6022b0e3cf5a2fa7b0173203648136063b314","access_level":"public","create_time":"$D2026-11-07T10:02:38.844Z","update_time":"$D2026-11-08T10:18:26.405Z","view_time":null,"first_public_time":"$D2026-11-09T05:32:11.359Z","has_unpublished_edits":null,"fork":false,"stars":1,"starred":false,"owner":{"type":"team","id":"6b95df6ccdd0cd40","login":"xoolive","name":"Xavier Olive","avatar_url":"https://avatars-next.observableusercontent.com/e055a445877c5a8ee2da134a257760536c2d1ee2fcd43f3aaf3f40db54e13733","tier":"starter_2024","bio":"Research scientist. I like maps, code and data visualisation. I keep an eye on anything that can fly, an ear for fipradio and a foot ready for next trail.","followed_by_you":false,"follows_you":false},"version":111,"publish_version":111,"role":null,"authored":false,"authors":[{"type":"individual","id":"0b2e8d88aeec5ab4","login":"xoolive","name":"Xavier Olive","avatar_url":"https://avatars-next.observableusercontent.com/e055a445877c5a8ee2da134a257760536c2d1ee2fcd43f3aaf3f40db54e13733","tier":null,"bio":"Research scientist. I like maps, code and data visualisation. I keep an eye on anything that can fly, an ear for fipradio and a foot ready for next trail.","followed_by_you":false,"follows_you":false}],"members":[{"type":"team","id":"6b95df6ccdd0cd40","login":"xoolive","name":"Xavier Olive","avatar_url":"https://avatars-next.observableusercontent.com/e055a445877c5a8ee2da134a257760536c2d1ee2fcd43f3aaf3f40db54e13733","tier":"starter_2024","role":"editor"}],"comments":[],"custom_thumbnail":false,"trashed":false,"lapsed":null,"subscribed":null,"forks":0,"collection":{"id":"a749676c7f0928cb","slug":"30-day-map-challenge","title":"30 day map challenge","owner":{"id":"6b95df6ccdd0cd40","login":"xoolive","name":"Xavier Olive"}},"type":"notebook","body":{"titling":"auto","title":"Day 9: Urban–rural","cells":[{"id":1,"value":"$${nav(9)}\n\n# Day 9: Urban–rural\n\nHow far is the nearest doctor? In France, the answer draws a line between town and country, though not quite the one you would expect. Each of the 35,000 communes is coloured by its *accessibilité potentielle localisée* (APL), the indicator the French ministry of health uses to spot *déserts médicaux*: how many consultations a year each inhabitant can get from the general practitioners within reach, counting their actual activity, the drive to their practice and the age of the population. In brown, the communes below 2.5 consultations a year, the threshold below which a commune counts as under-served. Hover over a commune for its figures; switch profession below.","pinned":false,"hidden":false,"mode":"md","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":2,"value":"// the five professions of the official APL\nviewof profession = Inputs.radio(new Map(Object.entries(apl.PROFESSIONS).filter(([, p]) => !p.recomputed).map(([key, p]) => [p.label, key])), {value: \"gp\", label: \"Profession\"})","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":3,"value":"apl.legend(d3, profession, national, Math.min(width, 420))","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":4,"value":"map = apl.drawMap(prepared, profession, national, d3)","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":5,"value":"## Not only a rural story\n\nShare of the population living in a commune below the threshold, by class of the Insee density grid, from the large cities to the most sparsely settled countryside. Medical deserts grow with distance from towns, but they are not only rural: the outer suburbs of Paris are among the largest. More than half of the inhabitants of the Val-d'Oise, the Essonne, the Seine-et-Marne and the Yvelines live below the threshold, in towns such as Argenteuil, Corbeil-Essonnes, Chelles or Sartrouville, while Marseille and its surroundings are well served.","pinned":false,"hidden":false,"mode":"md","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":6,"value":"Plot.plot({\n marginLeft: 130, marginRight: 20, height: 220, width: Math.min(width, 640),\n x: {percent: true, domain: [0, 100], label: \"Population below the threshold (%)\", grid: true},\n y: {domain: apl.DENSITY, label: null},\n marks: [\n Plot.barX(shares, {x: \"share\", y: \"label\", fill: \"#8c510a\"}),\n Plot.text(shares, {x: \"share\", y: \"label\", text: (d) => `${(100 * d.share).toFixed(0)} % of ${(d.pop / 1e6).toFixed(1)} M`, dx: 4, textAnchor: \"start\", fill: \"#444\"}),\n Plot.ruleX([0])\n ]\n})","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":7,"value":"$3c","pinned":false,"hidden":false,"mode":"md","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":16,"value":"## Beyond the official figures\n\nThe DREES computes the APL for five professions only. The same idea can be extended to others with the national directory of health professionals (RPPS), which lists every practitioner with their places of practice. The maps below are **not published figures**: they were recomputed for this page, with a cruder method explained underneath. Dermatologists and ophthalmologists are the specialists with the longest waits; speech therapists (*orthophonistes*) are the profession whose shortage is most talked about, for children above all.","pinned":false,"hidden":false,"mode":"md","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":17,"value":"// the layers recomputed from the RPPS directory\nviewof specialty = Inputs.radio(new Map(Object.entries(apl.PROFESSIONS).filter(([, p]) => p.recomputed).map(([key, p]) => [p.label, key])), {value: \"derm\", label: \"Profession\"})","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":18,"value":"apl.legend(d3, specialty, national2, Math.min(width, 420))","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":19,"value":"map2 = apl.drawMap(prepared, specialty, national2, d3)","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":20,"value":"Plot.plot({\n marginLeft: 130, marginRight: 20, height: 220, width: Math.min(width, 640),\n x: {percent: true, domain: [0, 100], label: \"Population below the threshold (%)\", grid: true},\n y: {domain: apl.DENSITY, label: null},\n marks: [\n Plot.barX(shares2, {x: \"share\", y: \"label\", fill: \"#8c510a\"}),\n Plot.text(shares2, {x: \"share\", y: \"label\", text: (d) => `${(100 * d.share).toFixed(0)} % of ${(d.pop / 1e6).toFixed(1)} M`, dx: 4, textAnchor: \"start\", fill: \"#444\"}),\n Plot.ruleX([0])\n ]\n})","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":21,"value":"$3d","pinned":false,"hidden":false,"mode":"md","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":8,"value":"Data: [APL 2024](https://data.drees.solidarites-sante.gouv.fr/explore/dataset/530_l-accessibilite-potentielle-localisee-apl/), DREES (SNIIR-AM, Cnam; populations and distances, Insee) · [Annuaire Santé, extraction RPPS](https://www.data.gouv.fr/datasets/annuaire-sante-extractions-des-donnees-en-libre-acces-des-professionnels-intervenant-dans-le-systeme-de-sante-rpps-format-zip), Agence du numérique en santé (dermatologists, ophthalmologists and speech therapists, 6 November 2026) · [Grille communale de densité 2024](https://www.insee.fr/fr/information/6439600), Insee · [Contours administratifs 2025](https://www.data.gouv.fr/datasets/contours-administratifs/), Etalab, from IGN Admin Express.","pinned":false,"hidden":false,"mode":"md","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":9,"value":"## Code","pinned":false,"hidden":false,"mode":"md","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":10,"value":"national = apl.stats(data.communes, profession)","pinned":true,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":11,"value":"shares = apl.byDensity(data.communes, profession, national.threshold)","pinned":true,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":12,"value":"prepared = apl.prepare(d3, data, Math.min(width, 960))","pinned":true,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":13,"value":"// Communes (APL of the five professions, homemade APL for dermatologists, ophthalmologists\n// and speech therapists, population, density class)\n// and départements, simplified TopoJSON\ndata = {\n const stream = (await FileAttachment(\"france-health-v3.json.gz\").stream()).pipeThrough(new DecompressionStream(\"gzip\"));\n const topo = await new Response(stream).json();\n const feature = (key) => topojson.feature(topo, topo.objects[key]).features.map((f) => apl.rewind(d3, f));\n return {communes: feature(\"communes-apl\"), departements: feature(\"departements\")};\n}","pinned":true,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":14,"value":"// Navigation header of the series\nimport {nav} from \"@xoolive/30-day-map-challenge-2026-day-1-points\"","pinned":true,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":15,"value":"$3e","pinned":true,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":22,"value":"national2 = apl.stats(data.communes, specialty)","pinned":true,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":23,"value":"shares2 = apl.byDensity(data.communes, specialty, national2.threshold)","pinned":true,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"}],"files":"$W3f","stdlib":"2"}} 33:null