// clean up process for benefits
viewof benefits = aq.from(clean)
.select('company', 'benefits') // select these columns only
.derive({
num_benefits: aq.escape( d => d3.count(d.benefits) ) // get a count of benefits by row
})
.dedupe('company','benefits') // deduplicate rows with the same benefit list
.filter(d => d.num_benefits > 0) // remove those without benefits (for other purposes, this might be useful)
// below is the part where I very ungracefully work with the arrays
// starts here
.derive({
benefits: d => op.replace(d.benefits, "['", "")
})
.derive({
benefits: d => op.replace(d.benefits, "']", "")
})
.derive({
benefits: d => op.replace(d.benefits, /'/g, "")
})
// ends here ^
.groupby('company') // once the array is clean, group all benefits by company
.rollup({
all_benefits: d => op.array_agg(d.benefits) // one company, one array of benefits
})
.derive({
all_benefits: d => op.split(d.all_benefits, ',')
})
.unroll('all_benefits') // do one company and one benefit per row
.derive({
all_benefits: d => op.trim(d.all_benefits)
})
.dedupe('company', 'all_benefits') // remove duplicates that resulted from the above
.rename({all_benefits: 'benefit'})
.view()
viewof unique_benefit_list = benefits
.select(1)
.dedupe()
.orderby('benefit')
.view()
// what companies offer benefits related to the benefits of interest?
viewof focus_benefits = benefits
// @Gretchen, you can add/remove benefits below if you'd like to explore others
.filter(d => d.benefit == 'Free or subsidised travel' || d.benefit == 'Company car')
.view()
// how many companies include the benefits in focus?
viewof focus_benefits_count = focus_benefits
.select('company')
.dedupe()
.count()
.view()
// how many of the companies that include the benefits in focus are part of the top companies?
focus_benefits
.filter(aq.escape(d => top_companies.includes(d.company)))
.view()
// nada :(
// what percent of the companies out of the whole dataset include the focused benefits?
viewof percent_benefits = benefits
.select('company')
.dedupe()
.count()
.rename({count: 'total number of companies'})
.derive({
percent_with_focus_benefits: aq.escape( d => focus_benefits_count.get('count',0) / d['total number of companies'] )
})
.view()
viewof clean_count_time_week = clean_count_overtime
.derive({date: d => op.parse_date(d.date)})
.derive({
week_date: aq.escape( d => d3.utcMonday(d.date) )
})
.groupby('week_date', 'company')
.rollup({
count: d => op.sum(d.count)
})
.groupby('week_date')
.derive({
percent: d => d.count/op.sum(d.count)
})
// .filter(d => +d.week_date == +op.utcdatetime(2023,1,27))
// .rollup({
// percent: d => op.sum(d.percent)
// })
.view()
top_companies = clean_count_overtime
.groupby('company')
.rollup({
count: d => op.sum(d.count)
})
.orderby(aq.desc('count'))
.filter(d => op.rank() <= 15)
.array('company')
palettes
viewof titles = aq.from(raw_data)
// .derive({
// date: d => op.parse_date(d.date)
// })
.select('job_title', 'company')
.dedupe()
.view()
import {us} from "@observablehq/build-your-first-choropleth-map-with-observable-plot"
states = topojson.feature(us, us.objects.states)
nation = topojson.feature(us, us.objects.nation)
counties_ = topojson.feature(us, us.objects.counties)
viewof clean_count_overtime = aq.from(raw_data)
.derive({date: d => op.parse_date(d.date)})
.groupby('date', 'company')
.count()
.orderby('date')
.view()
viewof clean = aq.from(raw_data)
.derive({date: d => op.parse_date(d.date)})
// .filter(d => d.company == 'Uber Drivers')
.view()
clean
.select('geo_city', 'geo_state')
.dedupe()
.view()
zip_codes = d3.csvParse(data)
prox_url = 'https://corsproxy.io?' + encodeURIComponent(v_data['zipcodes.csv'].url)
data = fetch(prox_url).then(response => response.text())
raw_data = file.csv({typed: false})
zip_map = new Map(zips.map(d => [d.zip_code, d]))
//broadband_access_2018 = new Map(us_broadband_2018.map(({id, broadband}) => [id, broadband]))
// this is a really cool way of displaying tables. Much more aesthetically pleasing than the defaults of arquero and input.table
import { formatTable } from "@saneef/pretty-tables"
date_parser = d3.utcParse('%m/%d/%Y')
appcast_color_palette = ['#3138BB', '#FF7D6F', '#AF9BFC', '#C7FFE2', '#127FFC', '#FF9B77']
// sample of how to place html blocks in a grid
// html `<div style="display: grid; grid-template-columns: 1fr 1fr 1fr; column-gap: 0px; row-gap: 5px;">
// ${viewof spend}
// ${viewof cpl}
// ${viewof cpc}
// </div>`
Purpose-built for displays of data
Observable is your go-to platform for exploring data and creating expressive data visualizations. Use reactive JavaScript notebooks for prototyping and a collaborative canvas for visual data exploration and dashboard creation.
