Senior Data Scientist, MassMutual
Lecturer, UC Berkeley MIDS
--
PhD Applied Math, Computational Story Lab, UVM
MS Applied Math, UVM
BS Math, Virginia Tech
Public
Edited
Nov 21
1 fork
chart = {
return embed(spec, {
renderer: "svg", // Explicitly use SVG renderer
config: {
// Use system fonts
text: {font: "system-ui, -apple-system, sans-serif"},
style: {
"guide-label": {font: "system-ui, -apple-system, sans-serif"},
"guide-title": {font: "system-ui, -apple-system, sans-serif"}
}
}
});
}
visWidth = width - 160
dataFilteredGender = data.filter(d => d.Gender === selectedGender)
stats = {
const currentAgeRow = dataFilteredGender.find(d => d.Year === selectedYear && d.Age === currentAge);
const meanAge = currentAge + currentAgeRow['e(x)'];
const survivalProbs = processedData.filter(d => d.Type === "Survival Probability");
const medianAge = survivalProbs.find(d => d.Probability <= 0.5)?.Age;
return {
meanAge,
medianAge,
currentAge
};
}
processedData = {
// Filter for selected year and age
const yearData = dataFilteredGender.filter(d => d.Year === selectedYear && d.Age >= currentAge);
// Calculate survival probabilities
let survivalProb = 1.0;
const probabilities = [];
yearData.forEach(row => {
probabilities.push({
Age: row.Age,
Probability: survivalProb,
Type: "Survival Probability"
});
survivalProb *= (1 - row['q(x)']);
});
// Calculate death probability density
const deathProbs = probabilities.map(d => d.Probability);
const ages = probabilities.map(d => d.Age);
const density = gradient(deathProbs.map(p => -p), ages);
const deathProbabilities = ages.map((age, i) => ({
Age: age,
Probability: density[i],
Type: "Death Probability Density"
}));
// Combine both sets of data
return [...probabilities, ...deathProbabilities];
}
// Function to calculate gradient
function gradient(values, domain) {
// Calculate central differences
const deltas = Array(values.length);
// Forward difference for first point
deltas[0] = (values[1] - values[0]) / (domain[1] - domain[0]);
// Central differences for middle points
for (let i = 1; i < values.length - 1; i++) {
deltas[i] = (values[i + 1] - values[i - 1]) / (domain[i + 1] - domain[i - 1]);
}
// Backward difference for last point
const last = values.length - 1;
deltas[last] = (values[last] - values[last - 1]) / (domain[last] - domain[last - 1]);
return deltas;
}
genders = {
return [...new Set(data.map(d => d.Gender))].sort();
}
years = {
return [...new Set(dataFilteredGender.map(d => d.Year))].sort();
}
visWidthMin = Math.max(visWidth, 500)
spec = {
const baseSpec = {
"width": visWidthMin,
"height": .4*visWidthMin,
"data": {"values": processedData},
"encoding": {
"x": {"field": "Age", "type": "quantitative", "title": "Age"}
},
"config": {
"legend": {"orient": "top", "layout": {"top": {"anchor": "middle"}}, "title": null}
},
"layer": [
{
// Survival probability line
"transform": [{"filter": "datum.Type === 'Survival Probability'"}],
"mark": "line",
"encoding": {
"y": {
"field": "Probability",
"type": "quantitative",
"title": "Survival Probability",
"axis": {"format": "%"}
},
"color": {
"field": "Type",
"type": "nominal",
"scale": {
"domain": ["Survival Probability", "Death Probability Density"],
"range": ["blue", "red"]
}
},
"tooltip": [
{"field": "Age", "type": "quantitative", "title": "Age"},
{"field": "Probability", "type": "quantitative", "title": "Probability", "format": ".1%"},
{"field": "Type", "type": "nominal"}
]
}
},
{
// Death probability density
"transform": [{"filter": "datum.Type === 'Death Probability Density'"}],
"mark": {"type": "area", "opacity": 0.3},
"encoding": {
"y": {
"field": "Probability",
"type": "quantitative",
"title": "Death Probability Density",
"axis": {"format": "%"},
"scale": {"zero": true}
},
"color": {"field": "Type", "type": "nominal"},
"tooltip": [
{"field": "Age", "type": "quantitative", "title": "Age"},
{"field": "Probability", "type": "quantitative", "title": "Probability", "format": ".2%"},
{"field": "Type", "type": "nominal"}
]
}
},
{
// Mean age line
"mark": {"type": "rule", "color": "#00AA00", "size": 2},
"encoding": {
"x": {"datum": stats.meanAge}
}
},
{
// Mean age label
"mark": {"type": "text", "align": "left", "dx": 5, "fill": "#00AA00"},
"encoding": {
"x": {"datum": stats.meanAge},
"y": {"value": 20}, // Position from top
"text": {"value": `Mean Age: ${stats.meanAge.toFixed(1)}`}
}
},
{
// Median age line
"mark": {"type": "rule", "color": "#FF4500", "size": 2},
"encoding": {
"x": {"datum": stats.medianAge}
}
},
{
// Median age label
"mark": {"type": "text", "align": "left", "dx": 5, "fill": "#FF4500"},
"encoding": {
"x": {"datum": stats.medianAge},
"y": {"value": 40}, // Position from top
"text": {"value": `Median Age: ${stats.medianAge.toFixed(1)}`}
}
}
],
"resolve": {
"scale": {
"y": "independent" // Make y-scales independent
}
}
};
return baseSpec;
}
data = FileAttachment("all.json").json()
embed = require("vega-embed@6")
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.