Maps — and other representations of the world.
Dotted OceansWorld Cities GabrielAntipodal ArchipelagoBi-Hexagonal ProjectionParis, TexasUSSR, 1938Hex NDVIIGRF-90🌖 Crater-shadingModified Stereographic Alaska GLSLVolcano on A-FrameBorn in AntarcticaNatural Earth issuesHello, spl.jsNéguentropieLoir-et-Cher isodémographiqueTheoretical land coverFive OceansPantelleria Joy PlotRouting dollarsDeriving PantelleriaGuessing CzechoslovakiaGHSL — Growing and decliningHistorical shapefilesThe E.N.D GameSpherical signed distanceWalk-on-Spheres… on a SpherePotential dollar flowsThirty days and as many maps
Ann Arbor pycnophylactic
tracts = Array.from(points, (p) => {
for (let i = 0; i < features.features.length; i++) {
if (d3.geoContains(features.features[i], p)) return i;
}
// try and repair the few holes that were in the original geojson file
for (let i = 0; i < features.features.length; i++) {
if (d3.geoContains(features.features[i], [p[0] + 0.001, p[1]])) return i;
}
for (let i = 0; i < features.features.length; i++) {
if (d3.geoContains(features.features[i], [p[0], p[1] + 0.001])) return i;
}
})
classes = d3.group(d3.range(points.length), (i) => tracts[i])
voronoi = d3.geoVoronoi(points)
total = features.features.map((d) => d.properties.CY7001)
counts = d3.rollup(
points,
(v) => v.length,
(p, i) => tracts[i]
)
function pycnoStep(values0) {
const values1 = Float64Array.from(values0);
const neighbors = voronoi.delaunay.neighbors;
for (const [tract, I] of classes) {
if (tract === undefined) continue;
// smooth by making our value closer to the mean of our neighbors
for (const i of I) {
values1[i] +=
(d3.mean(neighbors[i], (j) => values0[j]) - values0[i]) * 0.7;
}
// enforce sums
const sum = d3.sum(I, (i) => values1[i]);
for (const i of I) {
values1[i] += ((total[tract] - sum) / I.length) * 1;
}
// clamp each point above 0 and reallocate to others in the same class
const neg = d3.group(I, (i) => values1[i] < 0);
if (neg.has(true)) {
const reallocate =
-d3.sum(neg.get(true), (i) => values1[i]) / neg.get(false).length;
for (const i of neg.get(true)) values1[i] = 0;
for (const i of neg.get(false)) values1[i] += reallocate;
}
}
return values1;
}
// applies the reallocation *n* times, and updates the map
reallocated = {
let v = init;
for (let i = 0; i < steps; i++) {
v = pycnoStep(v);
}
pycnoMap.update(v);
return v;
}
error = Math.round(
d3.sum(
d3.rollup(
reallocated,
(v) => d3.sum(v),
(d, i) => tracts[i]
),
([tract, sum]) => Math.abs(total[tract] - sum)
)
)
points = Array.from(pick2d(width, height, 10000, gridtype), projection.invert)
color = d3
.scaleSequential(d3.interpolateCividis)
.domain(d3.extent(total.map((d, i) => d / classes.get(i).length)))
d3 = require("d3@7", "d3-geo-voronoi@1.6")
features = FileAttachment("annarbor.geojson").json()
projection = d3
.geoMercator()
.fitExtent([[20, 20], [width - 20, height - 20]], features)
height = 600
import { pick2d } from "@fil/2d-point-distributions"
One platform to build and deploy the best data apps
Experiment and prototype by building visualizations in live JavaScript notebooks. Collaborate with your team and decide which concepts to build out.
Use Observable Framework to build data apps locally. Use data loaders to build in any language or library, including Python, SQL, and R.
Seamlessly deploy to Observable. Test before you ship, use automatic deploy-on-commit, and ensure your projects are always up-to-date.