Published
Edited
May 28, 2020
17 stars
Bitemporal coronavirus casesSensitivity and Specificity (sketch)Coronavirus orbitsVarianceSmoothing with simple moving average vs. binomial filterCoronavirus logistic isoplethsBuilding up to John Burn-Murdoch’s Coronavirus trackerCovid Tracking Project data
Recreating John Burn-Murdoch’s Coronavirus tracker
Also listed in…
Coronavirus
chart = html`${style}
<svg class="covid" width="${width}" height="${height}">
<defs>
<clipPath id="${clip.id}">
<path d="
M ${x.range()[0]} ${y.range()[0]}
H ${x.range()[1]}
V ${y.range()[1]}
H ${x.range()[0]} Z
" />
</clipArea>
</defs>
<g transform="translate(${x.range()[0]}, ${y(threshold)})">
<text dy="1em" style="text-transform: uppercase; fill: gray;">
Days since ${threshold} cases
</text>
<path
d="M 0 0 H ${x.range()[1]}"
stroke="gray"
/>
</g>
<g style="text-transform: uppercase; fill: gray;">
<text x="${x.range()[0]}" y="${y.range()[1]}">Doubles every…</text>
${curves.map(
({ x0, T, data }) => svg`<g>
<path
d="${lineTrunc(data)}"
stroke-dasharray="2,2"
stroke="gray"
clip-path="${clip}"
/>
<text
x="${x(expSolveForX(x0, T)(y.domain()[1]))}"
y="${y.range()[1]}"
text-anchor="middle"
>
${T} days
</text>
</g>`
)}
</g>
<g clip-path="${clip}">
${series.filter(hasOffset).map(
({ key, latest, color, data, highlight, offset }) => svg`<g
opacity="${highlight ? 1 : .1}"
transform="translate(${-(x(offset) - x(0))}, 0)"
>
<path
d="${line(data.map(d => [d.i, d.value]))}"
stroke="white"
stroke-width="4"
/>
<path
d="${line(data.map(d => [d.i, d.value]))}"
stroke="${color}"
stroke-width="1.5"
/>
<text
x="${x(data.length - 1)}"
y="${y(latest)}"
fill="${color}"
dx=".2em"
dy=".3em"
font-weight="bold"
>
${key}
</text>
</g>`
)}
</g>
</svg>`
style = html`<style>
.covid path {
fill: none;
}
.covid text {
font-family: sans-serif;
font-size: 12px;
}
</style>`
// x0: initial value
// T: doubling time
exp = (x0, T) => t => x0 * 2 ** (t / T)
expSolveForX = (x0, T) => y => (T * Math.log(y / (x0 / 2))) / Math.LN2 - T
curves = d3.range(1, 8).map(T => ({
T,
x0: threshold,
data: range(0, domain.length, 1).map((d, i) => [i, exp(threshold, T)(d)])
}))
domain = Array.from(data.get(keys[0]).keys())
latest = domain[domain.length - 1]
keys = Array.from(data.keys())
series = keys.map(key => ({
key,
latest: data.get(key).get(latest),
color: hsl(hues.get(key)),
highlight: highlights.includes(key),
offset: xOffsets.get(key),
data: domain.map((d, i) => ({
date: dateParse(d),
i,
di: i - xOffsets.get(key),
value: data.get(key).get(d)
}))
}))
data = transformData({
statistic,
smoothing
})
xOffsets = mapEach(map =>
Array.from(map.values()).findIndex(d => d > threshold)
)(data)
hasOffset = ({ offset }) => offset !== -1
minOffset = d3.min(Array.from(xOffsets.values()).filter(d => d !== -1))
line = d3
.line()
.x(d => x(d[0]))
.y(d => y(d[1]))
.defined(d => logValid(d[1]))
lineTrunc = d => line(d.filter(([i, d]) => d < 20 * y.domain()[1]))
x.domain()
x = d3
.scaleLinear()
.domain([0, domain.length - minOffset + 1])
.range([margin.left, width - margin.right])
y = d3
.scaleLog()
.domain([threshold, max(data) * 1.1])
.range([height - margin.bottom, margin.top])
height = width / 2
max = map =>
d3.max(
d3
.merge(
Array.from(map, ([key, value]) =>
Array.from(value, ([key, value]) => value)
)
)
.filter(logValid)
)
hues = new Map(keys.map(d => [d, Math.floor(Math.random() * 360)]))
hsl = hue => `hsl(${hue}, 70%, 30%)`
valid = d => d !== null && d !== Infinity && d !== -Infinity && !isNaN(d)
logValid = d => valid(d) && d !== 0
margin = ({ top: 10, right: 40, bottom: 20, left: 10 })
clip = DOM.uid("clip")
// inclusive!
range = (min, max, step) => [...d3.range(min, max, step), max]
d3 = require("d3@5")
import { transformData, dateParse, stats } from "8961aecb496ce269"
import { mapEach } from "@tophtucker/scrapbook"
import { radio, slider, checkbox } from "@jashkenas/inputs"
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.