import { vl } from "@vega/vega-lite-api"
CH = d3.csvParse(await FileAttachment("CH.csv").text(), d3.autoType)
// Always start a vega-lite chart with "vl"
vl
// We can then chain vega-lite functions with a dot ".",
// starting with the mark:
.markCircle() // try .markBar() or .markPoint()
// use the dataset that we defined above:
.data(CH)
// invoking .render() draws the plot:
.render()
vl.markCircle()
.data(CH)
.encode(
// 1. Encode `elected` on the y axis
vl.y().field("elected").type("quantitative"),
// 2. Encode `year` on the x axis
vl.x().field("year").type("ordinal"), // type "ordinal" can be used here as years are ordered strings.
// 3. Use color to encode the dimension `gender`:
vl.color().fieldN("gender"), // shortcut for field & type definition: fieldN("gender")
)
.render()
vl.markArea() // Try using .markArea(), .markLine() or .markBar()
.data(CH)
.encode(
vl.x().fieldO("year"),
vl.y().fieldQ("elected"),
vl.color().fieldN("gender"),
// Try using additional encoding channel:
// vl.tooltip(["year", "elected", "gender"]),
// vl.facet().fieldN("gender"),
//vl.opacity().value(.6)
)
.render()
vl.markLine({
interpolate: "monotone" // "step" | "linear" | "monotone" | ...
})
.title("No gender parity in the Swiss National Council")
.width(400)
.height(200)
.data(CH)
.encode(
vl.x().fieldO("year")
.title('Year of election'),
vl.y().fieldQ("elected")
.title('Number of MP in the National Council')
.axis({ tickCount: 2 }),
vl.color().fieldN("gender")
.scale({ domain: ["Frau", "Mann"], range: ["orange", "gray"] })
.title(null)
.legend({ orient: "top" }),
vl.strokeWidth().value(3),
// vl.strokeDash().fieldN("gender")
)
.render(
{renderer: "SVG" }
)
vl.markArea()
.data(CH)
// Let's focus on the number for women in the Council:
.transform(
vl.filter("datum.gender === 'Frau' ")
)
.encode(
vl.x().fieldO("year"),
vl.y().fieldQ("elected"),
vl.color().fieldN("gender")
.scale({ domain: ["Frau", "Mann"], range: ["orange", "gray"] })
)
.render()
vl.markBar()
.data(CH)
.transform(
vl.groupby("gender").aggregate(
vl.average("elected").as("gender_aggregated")) // Try median or sum
)
.encode(
vl.x().fieldN("gender"),
vl.y().fieldQ("gender_aggregated"),
vl.tooltip(["gender_aggregated"]),
vl.color().fieldN("gender")
.scale({ domain: ["Frau", "Mann"], range: ["orange", "gray"] })
)
.render()
vl.markPoint({ size: 70, color: "black" })
.data(CH)
.transform(
vl.groupby("year")
.pivot("gender")
.value("elected")
)
.encode(
// Two new field names added by the transformation
vl.x().fieldQ("Frau")
.axis({ titleColor: "orange", titleFontSize: 16 }),
vl.y().fieldQ("Mann")
.axis({ titleColor: "gray", titleFontSize: 16 }),
)
.render()
vl.markArea()
.encode(
vl.x().fieldN("year"),
vl.y().fieldQ("elected"),
vl.color().fieldN("gender")
.scale({ domain: ["Frau", "Mann"], range: ["orange", "gray"] }),
vl.tooltip(["year", "gender", "elected"])
)
.width(500)
.height(100)
// We can create small multiples using "facets":
.facet({ row: vl.field("party") })
.data(parties)
.render()
{
const points = vl.markCircle()
.encode(
vl.x().fieldQ("proportion_of_women")
.sort(vl.field("year").order("ascending")),
vl.y().fieldQ("number_of_seats")
.sort(vl.field("year").order("ascending")),
vl.color().fieldN("party")
.scale({ domain: ["CVP", "FDP", "SP", "SVP"], range: ['#D6862B', '#3872B5','#F0554D','#4B8A3E']}),
// Note the use of a conditional circle size to highlight 2019:
vl.size().if("datum.year === 2019", vl.value(100)).value(30),
vl.fillOpacity().value(2),
vl.tooltip(["year", "party", "proportion_of_women"])
)
.width(150)
.height(150)
const lines = points.markLine({ interpolate: "monotone" })
.encode(
vl.size().value(1)
)
// We can layer points on top of lines using vl.layer():
return vl.layer(points, lines)
.facet(vl.field("party"))
.data(parties)
.transform(
vl.groupby(["year", "party"]).pivot("gender").value("elected"),
vl.calculate("datum.Frau + datum.Mann").as("number_of_seats"),
vl.calculate("(datum.Frau / datum.number_of_seats) * 100").as("proportion_of_women")
)
.render({renderer: "SVG"});
}
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.