7.9 KiB
Animated treemap
This treemap of U.S. population uses d3.treemapResquarify for a stable layout with changing node values. Data: U.S. Census Bureau
const scrubber = Scrubber(d3.range(data.keys.length), {
delay: 2500,
loop: false,
format: (i) => data.keys[i]
});
const index = view(scrubber);
const chart = display(AnimatedTreemap());
The chart is created by the AnimatedTreemap function:
function AnimatedTreemap() {
const width = 928;
const height = width;
// This is normally zero, but could be non-zero if this cell is
// re-evaluated after the animation plays.
const initialIndex = scrubber.value;
// To allow the transition to be interrupted and resumed, we parse
// the displayed text (the state population) to get the current
// value at the start of each transition; parseNumber and
// formatNumber must be symmetric.
const parseNumber = (string) => +string.replace(/,/g, "");
const formatNumber = d3.format(",d");
// Get the maximum total population across the dataset. (We know
// for this dataset that it’s always the last value, but that isn’t
// true in general.) This allows us to scale the rectangles for
// each state to be proportional to the max total.
const max = d3.max(data.keys, (d, i) => d3.hierarchy(data.group).sum((d) => d.values[i]).value);
// The category10 color scheme per state, but faded so that the
// text labels are more easily read.
const color = d3.scaleOrdinal()
.domain(data.group.keys())
.range(d3.schemeCategory10.map((d) => d3.interpolateRgb(d, "white")(0.5)));
// Construct the treemap layout.
const treemap = d3.treemap()
.size([width, height])
.tile(d3.treemapResquarify) // to preserve orientation when animating
.padding((d) => (d.height === 1 ? 1 : 0)) // only pad parents of leaves
.round(true);
// Compute the structure using the average value (since this
// orientation will be preserved using resquarify across the
// entire animation).
const root = treemap(d3.hierarchy(data.group)
.sum((d) => (Array.isArray(d.values) ? d3.sum(d.values) : 0))
.sort((a, b) => b.value - a.value));
const svg = d3.create("svg")
.attr("width", width)
.attr("height", height + 20)
.attr("viewBox", [0, -20, width, height + 20])
.attr("style", "max-width: 100%; height: auto; font: 10px sans-serif; overflow: visible;");
// Draw a box representing the total population for each time. Only
// show the boxes after the current time (to avoid distracting gray
// lines in between the padded treemap cells).
const box = svg.append("g")
.selectAll("g")
.data(data.keys.map((key, i) => {
const value = root.sum((d) => d.values[i]).value;
return { key, value, i, k: Math.sqrt(value / max) };
})
.reverse())
.join("g")
.attr("transform", ({k}) => `translate(${((1 - k) / 2) * width},${((1 - k) / 2) * height})`)
.attr("opacity", ({i}) => (i >= initialIndex ? 1 : 0))
.call((g) => g.append("text")
.attr("y", -6)
.attr("fill", "#777")
.selectAll("tspan")
.data(({ key, value }) => [key, ` ${formatNumber(value)}`])
.join("tspan")
.attr("font-weight", (d, i) => (i === 0 ? "bold" : null))
.text((d) => d))
.call((g) => g.append("rect")
.attr("fill", "none")
.attr("stroke", "#ccc")
.attr("width", ({ k }) => k * width)
.attr("height", ({ k }) => k * height));
// Render the leaf nodes of the treemap.
const leaf = svg.append("g")
.selectAll("g")
.data(layout(initialIndex))
.join("g")
.attr("transform", (d) => `translate(${d.x0},${d.y0})`);
leaf.append("rect")
.attr("id", (d) => (d.leafUid = uid("leaf")).id)
.attr("fill", (d) => { while (d.depth > 1) d = d.parent; return color(d.data[0]); })
.attr("width", (d) => d.x1 - d.x0)
.attr("height", (d) => d.y1 - d.y0);
// Clip the text to the containing node.
leaf.append("clipPath")
.attr("id", (d) => (d.clipUid = uid("clip")).id)
.append("use")
.attr("xlink:href", (d) => d.leafUid.href);
// Generate two tspans for two lines of text (name and value).
leaf.append("text")
.attr("clip-path", (d) => d.clipUid)
.selectAll("tspan")
.data((d) => [d.data.name, formatNumber(d.value)])
.join("tspan")
.attr("x", 3)
.attr("y", (d, i, nodes) => `${(i === nodes.length - 1) * 0.3 + 1.1 + i * 0.9}em`)
.attr("fill-opacity", (d, i, nodes) => (i === nodes.length - 1 ? 0.7 : null))
.text((d) => d);
leaf.append("title")
.text((d) => d.data.name);
// Scale the treemap layout to fit within a centered box whose area
// is proportional to the total current value. This makes the areas
// of each state proportional for the entire animation.
function layout(index) {
const k = Math.sqrt(root.sum((d) => d.values[index]).value / max);
const tx = ((1 - k) / 2) * width;
const ty = ((1 - k) / 2) * height;
return treemap.size([width * k, height * k])(root)
.each((d) => ((d.x0 += tx), (d.x1 += tx), (d.y0 += ty), (d.y1 += ty)))
.leaves();
}
// Expose an update method on the chart that allows the caller to
// initiate a transition. The given index represents the frame
// number (0 for the first frame, 1 for the second, etc.).
return Object.assign(svg.node(), {
update(index, duration) {
box.transition()
.duration(duration)
.attr("opacity", ({ i }) => (i >= index ? 1 : 0));
leaf.data(layout(index))
.transition()
.duration(duration)
.ease(d3.easeLinear)
.attr("transform", (d) => `translate(${d.x0},${d.y0})`)
.call((leaf) => leaf.select("rect")
.attr("width", (d) => d.x1 - d.x0)
.attr("height", (d) => d.y1 - d.y0))
.call((leaf) => leaf.select("text tspan:last-child")
.tween("text", function (d) {
const i = d3.interpolate(parseNumber(this.textContent), d.value);
return function (t) { this.textContent = formatNumber(i(t)); };
}));
},
});
}
This triggers the animation from the scrubber:
const update = chart.update(index, 2500);
The AnimatedTreemap function consumes a specific data structure, that lists the years as keys and a hierarchy of regions and states with an array describing the population (according to the census) of each State for every key:
display(data);
The source data is loaded from two file attachments—one in CSV format and the other in TSV format. For your own chart you’ll want to create a similar data structure—maybe by reading from an API with d3.csv, or by running a sql query on a database.
const keys = d3.range(1790, 2000, 10);
const [regions, states] = await Promise.all([
FileAttachment("/data/census-regions.csv").csv(), // for grouping states hierarchically
FileAttachment("/data/population.tsv").tsv(), // a wide dataset of state populations over time
]).then(([regions, states]) => [
regions,
states.slice(1).map((d) => ({
name: d[""], // the state name
values: keys.map((key) => +d[key].replace(/,/g, "")) // parse comma-separated numbers
}))
]);
const regionByState = new Map(regions.map((d) => [d.State, d.Region]));
const divisionByState = new Map(regions.map((d) => [d.State, d.Division]));
const data = {
keys,
group: d3.group(
states,
(d) => regionByState.get(d.name),
(d) => divisionByState.get(d.name)
)
};
These utilities help create the scrubber, and unique ids.
import {Scrubber} from "/components/scrubber.js";
import {uid} from "/components/DOM.js";