Digital Humanist. Manhattan-based.
Published
Edited
May 10, 2022
2 forks
7 stars
Plot: A few custom time axesAccess to Family planningPlot: MarimekkoAutoplot matrixWhat do people buy on Bandcamp?Plot: colorContrast transformA Timeline of Democratic Presidential CandidatesPlot: Diverging stacked barsPlot: Horizon ChartPlot: Ridgeline PlotTwo-Tone Pseudo Color Scales with Observable Plot & Vega-LiteRecreating Östling’s regression visualizationsError barsPlot: regressionSkies PlotHorizon graph & Barcode with PlotTemporal AliasingPlot for mathematiciansPlot Isotype dot plotCOVID-19 Netherlands Reproduction Number TutorialPlot Mountain SunsetSpine chartsGreenhouse gas emission projectionsRaincloud Plots with Observable PlotUpset Plots with Observable PlotExploring CIELChabNature variant waves graphicWhen Presidents Fade AwayOur World In Data charts visualisationThe normal modelVisualizing The New York Times’s Mini CrosswordDistributions and summary statistics - a collection of Plot examplesVariants of SARS-Cov-2 in EuropeLaid off: 6 million jobs and countingConcentration values vs. TimeTopological subsamplingPlot of plotsCyclical time scale, v2Bullet graph IIPlot: Bullet graphCyclical time scaleRoad traffic under COVID restrictions (France)K-means binningMatrix diagramELD ViewerHappy anniversary, Project Gutenberg!ROC curvePlot stargazer historyPlot HyperboloidPhone bar plotsLog-scale histogramDatabase Outage PlotBertin’s hotelIreland’s top trade partnersBump chartBubble MatrixRenko ChartGitHub BurndownPlot: Grid choroplethCandlestick Chart
The Coronavirus landscape
Also listed in…
Visualization
Type JavaScript, then Shift-Enter. Ctrl-space for more options. Arrow ↑/↓ to switch modes.
p = Plot.plot({
caption: html`Data from Johns Hopkins: chart by Ben Schmidt:<br>https://observablehq.com/@bmschmidt/the-coronavirus-landscape`,
height: places.size * 25,
width: width,
color: {
scheme: "dark2",
domain: ordering.filter(d => places.has(d)),
},
// marginTop: 50,
// marginBottom: 50,
x: {
tickFormat: "%b",
grid: true
},
y: {
position: "right",
range: [20, -80],
axis: null
},
fy: {
domain: ordering.filter(d => places.has(d)),
//padding: 0,
round: true,
axis: null,
display: false,
"fontSize": 0
},
facet: {
data: smooth_filtered,
y: "place",
marginLeft: 30,
marginRight: 150
},
marks: [
Plot.ruleY([0])
,
Plot.dot(smooth_filtered, {
filter: d => '' + d.date == cap_date_string,
r: 5,
x: "date",
y: 0,
dy: -4,
dx: 3,
fillOpacity: .5,
fy: "place",
fill: "place",
// curve: "catmull-rom"
}),
Plot.text(smooth_filtered, {
filter: d => '' + d.date == cap_date_string,
r: 5,
x: "date",
y: 0,
dy: -3,
textAnchor: 'start',
dx: 7,
fontSize: 14,
fy: "place",
text: "place",
// curv
}),
Plot.areaY(smooth_filtered, {
x: "date",
y: quantity,
fy: "place",
fill: "place",
title: "place",
fillOpacity: 0.5,
// curve: "catmull-rom"
}),
Plot.line(smooth_filtered, {
x: "date",
fill: "avg",
fy: "place",
y: quantity,
title: quantity
// rx: 20 // uncomment for circles
}),
]
})
smooth_filtered = smoothed.filter(d => d.date <= cap_date)
cap_date = new Date(maxdate)
cap_date_string = '' + cap_date
maxdate = "" + d3.max(smoothed, d => d.date)
ordering = {
if (dataset == "States") return linear_us_state_order
if (dataset == "Countries") return root_ordering
if (dataset == "US Metros") return metro_ordering
throw "Bad dataset"
}
import {Scrubber} from "@mbostock/scrubber"
async function insert_json(name, array, client) {
await client.query(`DROP TABLE IF EXISTS ${name}`)
const texted = d3.csvFormat(array)
await client.insertCSV(name, new TextEncoder("utf-8").encode(texted))
return
}
fips_msa = {
client
const fips_msa = await fetch(`https://gist.githubusercontent.com/bmschmidt/62ed8da02750cbf0b75278cfe06f7bb7/raw/37cc7dfe54b71c041fbc2d3d9cd214443d206c3c/msa-csa.json`).then(d => d.json())
const rows = [...Object.entries(fips_msa)].map(([fips, msa]) => {
const metro_state = msa.split(", ").pop().split("-")[0]
return {fips, msa, metro_state}
})
const texted = d3.csvFormat(rows)
await insert_json("fips_msa", rows, client)
return client.table('SELECT * FROM fips_msa')
}
viewof metro_values = {
fips_msa
return client.table(`SELECT metro_state, msa as name, SUM(population) as pop FROM fips_msa NATURAL JOIN county_populations GROUP BY metro_state, msa`)
}
metro_values[0]["metro_state"]
metro_ordering = {
metro_values.sort((a, b) => linear_us_state_order.indexOf(a["metro_state"]) - linear_us_state_order.indexOf(b["metro_state"]))
return metro_values.map(d => d["name"])
}
d3 = require('d3-array', 'd3-fetch', 'd3-dsv')
import {DuckDBClient} from '@cmudig/duckdb'
datenames = dates.map(d => {const [a, b, c] = d.split("/"); return `20${c}-${a}-${b}`})
client.table("DESCRIBE global")
dates = client.query("DESCRIBE global").then(d => d.getChild("column_name").toArray().slice(4))
client = {
const client = new DuckDBClient()
const db = await client.db()
await client.insertCSV("global", r)
await client.insertCSV("us", us_raw)
return client
}
places = new Set(smoothed.map(d => d.place).filter(d => ordering.indexOf(d) > -1))
mindate = new Date("2020-03-05")
smoothed = smoothed_unfiltered.filter(d => new Set(ordering).has(d.place)).filter(d => d.date > mindate)//.filter(d=>d.place.match("New ") )
viewof smoothed_unfiltered = {
wide
await client.query(`DROP TABLE IF EXISTS pops`)
await client.query(`CREATE TABLE pops (place CHAR, population INT)`)
let pop_set;
if (dataset === "States") {
pop_set = state_pops
} else if (dataset == "Countries") {
pop_set = populations
} else {
pop_set = metro_values.map( d=> {
return {
name : d.get("name"),
"population_total" : d.get("pop")
}
})
}
let popmin = 0
if (dataset == "Countries") popmin = 10000000
if (dataset == "US Metros") popmin = 150000
const pop_vals = pop_set.map(({name, population_total}) => `('${name.replace("'", "")}', ${population_total})`).join(", ")
await client.query(`INSERT INTO pops VALUES ${pop_vals}`)
return client.table(`SELECT *, CASE WHEN smoothed > 0 THEN smoothed*1000000/population ELSE 0 END as avg FROM ( SELECT *, (AVG(new_cases) OVER (
PARTITION BY place
ORDER BY date
RANGE BETWEEN INTERVAL 13 DAYS PRECEDING
AND INTERVAL 0 DAYS FOLLOWING
)) as smoothed FROM tmp NATURAL JOIN pops WHERE population > ${popmin} order by place, date) t1 `)
}
d = client.db()
metros = {
wide;
fips_msa;
await client.query(`DROP TABLE IF EXISTS metro_long`)
// Create metro counts
await client.query(`CREATE TABLE metro_long AS SELECT
msa place, date, sum(cases) as cases FROM us_long_raw NATURAL JOIN "fips_msa" GROUP BY place, date`)
return client.table(`SELECT * FROM metro_long`)
}
wide = {
await client.query(`DROP TABLE IF EXISTS long`)
await client.query(`DROP TABLE IF EXISTS us_long`)
await client.query(`DROP TABLE IF EXISTS us_long_raw`)
await client.query(`
CREATE TABLE IF NOT EXISTS long AS
SELECT "place", date, SUM(cases) cases FROM (
SELECT
"Country/Region" place,
unnest(${JSON.stringify(datenames).replaceAll('"', "'")})::DATE as date,
unnest(${JSON.stringify(dates)}) as cases
FROM global)
as t1
GROUP BY place, "date" ORDER BY place, date
`)
await client.query(`
CREATE TABLE us_long_raw AS
SELECT
"Province_State", FIPS,
unnest(${JSON.stringify(datenames).replaceAll('"', "'")})::DATE as date,
unnest(${JSON.stringify(dates)}) as cases
FROM us
`)
await client.query(`
CREATE TABLE us_long AS SELECT "Province_State" place, date, SUM(cases) cases
FROM "us_long_raw" GROUP BY place, date ORDER BY place, date`)
}
{
wide
let upper_tab = "long"
if (dataset == "States") upper_tab = "us_long"
if (dataset == "US Metros") upper_tab = "metro_long"
await client.query(`DROP TABLE IF EXISTS tmp`)
await client.query(`CREATE TABLE tmp AS
SELECT cases::FLOAT - LAG(cases, 1) OVER
(PARTITION BY 'place' ORDER BY 'date')::FLOAT as new_cases,
*
FROM ${upper_tab}
`)
return client.table(`SELECT * FROM tmp LIMIT 30`)
}
r = fetch(`https://raw.githubusercontent.com/CSSEGISandData/COVID-19/master/csse_covid_19_data/csse_covid_19_time_series/time_series_covid19_confirmed_global.csv`).then(d => d.arrayBuffer())
client.query(`CREATE TABLE
us_raw = fetch(`https://raw.githubusercontent.com/CSSEGISandData/COVID-19/master/csse_covid_19_data/csse_covid_19_time_series/time_series_covid19_confirmed_US.csv`).then(d => d.arrayBuffer())
import { linear_us_state_order } from '@bmschmidt/useful-linear-orders-for-countries-and-states'
import { populations, root_ordering } from '@bmschmidt/well-ordered-coronavirus-heatmaps-for-us-and-the-world'
root_ordering
import { state_info } from "@bmschmidt/state-info"
state_info.get("Alabama")
county_populations =
{
client;
const data = await d3.json('https://raw.githubusercontent.com/Zoooook/CoronavirusTimelapse/master/static/population.json')
for (let datum of data) {
datum.population = +datum.population
datum.fips = +datum.us_county_fips
}
await insert_json("county_populations", data, client)
return
}
state_pops = Object.entries(population.states).map(([state, pop]) => ({name: state, population_total: pop}))
import { population } from "@youbastard/us-state-populations"
[...state_info.entries()][3]
import { duckdb, db } from '@cmudig/duckdb'
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.