Public
Edited
Feb 29, 2024
Importers
DEFAULT_N = 10000
import {
dist,
distValue,
kde,
to,
exponential,
normalInterval
} with { DEFAULT_N } from "@huw/dist"
viewof selfGuidedDepressionSMD = dist`${normalInterval(0.15, 0.31, {
p: 0.95
})}`
viewof socialDesirabilityDiscount = dist`0.7 to 0.9`
viewof publicationDiscount = dist`0.75 to 0.9`
viewof scaleDiscount = dist`0.7 to 0.85`
viewof effectSizeAdjustment = Inputs.range([0.1, 2.0], {
label: "Effect size adjustment",
value: 1
})
viewof selfGuidedDepressionSMDDiscounted = dist`${selfGuidedDepressionSMD} * ${socialDesirabilityDiscount} * ${publicationDiscount} * ${scaleDiscount} * ${effectSizeAdjustment}`
viewof expertLifeSatisfactionStdev = dist`(1.9 to 2.7) * 0.6 + 1.86 * 0.4`
viewof depressionToLifeSatisfaction = dist`0.7 to 1`
viewof selfGuidedWellby = dist`${selfGuidedDepressionSMDDiscounted} * ${expertLifeSatisfactionStdev} * ${depressionToLifeSatisfaction}`
viewof retentionAdjustment = Inputs.range([0.1, 2.0], {
label: "Retention adjustment",
value: 1
})
viewof mentalHealthAppRetention = dist`${exponential(
1 / 0.04
)} * ${retentionAdjustment}`
viewof perUserWellby = dist`${selfGuidedWellby} * ${mentalHealthAppRetention}`
viewof acquisitionCostAdjustment = Inputs.range([0.1, 2], { label: "CPI adjustment", value: 1 })
viewof marketingCostPerDownload = dist`(0.5 to 6) * ${acquisitionCostAdjustment}`
{
const data = [
{ Name: "Headspace", Employees: 1200, Downloads: 70_000_000 },
{ Name: "Calm", Employees: 400, Downloads: 135_000_000 },
{ Name: "Breeze", Employees: 20, Downloads: 2_000_000 },
{ Name: "BetterHelp", Employees: 5000, Downloads: 80_000_000 },
{ Name: "Paired", Employees: 250, Downloads: 8_000_000 },
{ Name: "Insight Timer", Employees: 350, Downloads: 10_000_000 },
{ Name: "Balance", Employees: 57, Downloads: 8_000_000 },
{ Name: "Shmoody", Employees: 5, Downloads: 1_000_000 },
{ Name: "Daylio", Employees: 5, Downloads: 20_000_000 },
{ Name: "Bearable", Employees: 7, Downloads: 600_000 }
];
return Plot.plot({
x: { type: "log", grid: true },
y: { type: "log", grid: true },
marks: [
Plot.dot(data, {
x: "Employees",
y: "Downloads",
fill: "black",
channels: { Name: "Name" },
tip: true
}),
Plot.linearRegressionY(data, { x: "Employees", y: "Downloads" })
]
});
}
viewof downloadsPerEmployee = dist`25000 to 2500000`
viewof employeesPerDownloadAdjustment = Inputs.range([0.1, 2], { label: "Employees per download adjustment", value: 1 })
viewof employeesPerDownload = dist`(1 / ${downloadsPerEmployee}) * ${employeesPerDownloadAdjustment}`
viewof costPerEmployeeAdjustment = Inputs.range([0.1, 2], {
label: "Cost per employee adjustment",
value: 1
})
viewof costPerEmployee = dist`(25000 to 250000) * ${costPerEmployeeAdjustment}`
viewof developmentCostPerDownload = dist`${employeesPerDownload} * ${costPerEmployee}`
viewof totalCostPerDownload = dist`(${marketingCostPerDownload} + ${developmentCostPerDownload}) * (1.1 to 1.5)`
mentalHealthAppCostEffectiveness = distValue`${perUserWellby} / ${totalCostPerDownload} * 1000`
plot = {
const density = [...kde.density1d(mentalHealthAppCostEffectiveness)];
const nearestQuantilesInDensity = [0.5]
.map((p) => d3.quantile(mentalHealthAppCostEffectiveness, p))
.map((q) => ({
i: d3.bisectCenter(
density.map((d) => d.x),
q
),
q
}))
.map(({ q, i }) => ({ x: q, y: density[i].y }));
return Plot.plot({
y: { ticks: 0, label: "Density Estimate" },
x: {
domain: [
d3.min(mentalHealthAppCostEffectiveness),
d3.max([d3.quantile(mentalHealthAppCostEffectiveness, 0.97), 70])
],
label: "Cost-Effectivess (WELLBYs/US$1000)"
},
color: { legend: true },
height: 256,
width,
marks: [
Plot.areaY(density, { x: "x", y: "y", fillOpacity: 0.1 }),
Plot.lineY(density, { x: "x", y: "y" }),
Plot.ruleY([0]),
Plot.ruleX(
[
{ Name: "GiveDirectly", "Cost-Effectiveness": 8 },
{ Name: "AMF", "Cost-Effectiveness": 70 },
{ Name: "StrongMinds", "Cost-Effectiveness": 17 }
],
{ x: "Cost-Effectiveness", stroke: "Name", tip: true }
),
Plot.ruleX(nearestQuantilesInDensity, {
x: "x",
y2: "y",
strokeDasharray: [5, 3]
}),
Plot.axisX(),
Plot.axisX(
nearestQuantilesInDensity.map(({ x }) => x),
{ textStroke: "#fff", textStrokeWidth: 5 }
),
Plot.crosshairX(density, { x: "x", y: "y" })
]
});
}
dist`${mentalHealthAppCostEffectiveness}`
Inputs.bind(
Inputs.range([0.1, 2], { label: "Hiring adjustment" }),
viewof employeesPerDownloadAdjustment
)
Inputs.bind(Inputs.range([0.1, 2], { label: "Salary adjustment" }), viewof costPerEmployeeAdjustment)
Inputs.bind(Inputs.range([0.1, 2], { label: "CPI adjustment" }), viewof acquisitionCostAdjustment)
Inputs.bind(
Inputs.range([0.1, 2], { label: "Retention adjustment" }),
viewof retentionAdjustment
)
Inputs.bind(
Inputs.range([0.1, 2], { label: "Effect size adjustment" }),
viewof effectSizeAdjustment
)
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.