data analyst • mediapart
Published
Edited
Jun 15, 2021
import {sondages, desc as descNspPolls} from '@taniki/regionales2021-donnees-nsppolls'
confiance = 0.95
Inputs.table(t1Hypothese.tetes_liste)
liste = t1Hypothese.tetes_liste[1]
Z = jStat.normal.inv((1+confiance)/2, 0, 1)
${ t1Hypothese
.tetes_liste
.slice()
.sort((a,b) => b.intentions - a.intentions)
.map(liste => t1ListGraph(liste, t1Hypothese.sous_echantillon))
}
t1ListGraph = (liste, n) => {
let N = listN(liste.intentions/100, n)
let ic = [N.inv(0.025), N.inv(1-(1-confiance)/2)] // pour vérification
//return [p, s, ic]
return Plot.plot({
caption: `densité de probabilité d'intention de vote pour la liste de ${liste.tete_liste} (${liste.parti})`,
height: 200,
x: {
domain: [0, 50]
},
y:{
domain: [0,50]
},
marks: [
Plot.ruleX([5], { stroke: 'red'}),
Plot.ruleX([10], { stroke: 'green'}),
Plot.areaY(d3.range(0, 5, 0.1), { x: d => d, y: d => N.pdf(d/100), fill: 'red' }),
Plot.areaY(d3.range(5, 10, 0.1), { x: d => d, y: d => N.pdf(d/100), fill: '#999' }),
Plot.areaY(d3.range(10,100, 0.1), { x: d => d, y: d => N.pdf(d/100), fill: 'green'})
]
})
}
t1ListRidgeline(t1Hypothese, 'premier tour')
t1ListRidgeline = (hypothese, tour, row = 25) => {
let yStep = 15
return Plot.plot({
caption: html`
densité de probabilité d'intentions de vote au ${ tour } des élections régionales 2021<br>
données : <a href="https://twitter.com/nsppolls">NspPolls</a> · <a href="${ sondage.lien }">${ sondage.nom_institut }</a>`,
height: hypothese.tetes_liste.length * row,
marginLeft: 150,
x: {
domain: [0, 50],
label: 'intentions de vote (%)',
grid: true,
ticks: 10,
},
y:{
axis: false,
domain: [0, hypothese.tetes_liste.length * row]
},
marks: [
Plot.ruleX([5], { stroke: 'red', strokeOpacity: 0.3}),
Plot.ruleX([10], { stroke: 'green', strokeOpacity: 0.3}),
hypothese.tetes_liste.slice().sort((a,b)=>b.intentions - a.intentions).map((liste,i) => {
let N = listN(liste.intentions/100, hypothese.sous_echantillon)
let y0 = i*yStep
return [
Plot.area(d3.range(0, 5, 0.1), { x1: d => d, x2: d => d, y1: d => y0, y2: d => y0+N.pdf(d/100), fill: 'red', fillOpacity: 0.5 }),
Plot.area(d3.range(5, 10, 0.1), { x1: d => d, x2: d => d, y1: d => y0, y2: d => y0+N.pdf(d/100), fill: '#999', fillOpacity: 0.5 }),
Plot.area(d3.range(10,100, 0.1), { x1: d => d, x2: d => d, y1: d => y0, y2: d => y0+N.pdf(d/100), fill: 'green', fillOpacity: 0.5}),
Plot.text([[i, liste.tete_liste]], {
text: d => d[1],
textAnchor: 'end',
dy: 0,
y: d => y0,
x: d => N.inv(0.001) * 100
})
]
})
]
})
}
listN = (p,n) => {
//let p = liste.intentions / 100
let s = p * (1-p) / n
let N = jStat.normal(p, Math.sqrt(s,2))
let ic = [N.inv(0.025), N.inv(1-(1-confiance)/2)] // pour vérification
return N
}
import { eDiff } from '@taniki/regionales2021-ecarts-checker'
jStat = require('jStat')
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.