Berapa Jam Kerja

How many hours of minimum-wage work do rice, a room, a bus ride or an iPhone cost in Indonesia, compared with Southeast Asia, Korea and Japan? Every price on the site is converted into work time.

Status: all sections built: compare two cities, map, one item across cities, wishlist, monthly basket, regions inside Indonesia, Jakarta since 2018, salary calculator, "Kamu di persentil berapa?" (income percentile, see METHOD.md), context, editable data table. Plus /status.html, which shows whether every number is still current. Bahasa Indonesia by default, English toggle, light/dark theme. Lighthouse mobile (local build): performance 92–93, accessibility 100, best practices 100, SEO 100.

Run it

npm install
npm run dev            # local site
npm run check-data     # validate /data (also runs as part of build)
npm test               # unit tests for the method
npm run build          # check-data + typecheck + static build into dist/

The site builds for the domain root. BASE=/<path>/ npm run build builds for a sub-path instead (e.g. GitHub Pages).

Deploy

Live at https://indonesia-tanggal-tua.vercel.app/. Vercel builds and publishes every push to main (and gives each pull request a preview link); its build runs npm run build, so a data error stops the deploy. .github/workflows/check.yml runs tests and the build on pull requests.

Share images

npm run og screenshots each shareable view into public/og/*.png and writes public/s/<view>.html (Open Graph tags + redirect to #<view>). It needs Playwright (npm i -D playwright && npx playwright install chromium). Re-run after data changes that the share cards show, and commit the output.

Project layout

data/                     all numbers, versioned; nothing numeric lives in code
  places.json             cities: wage, wage type, monthly hours, FX, rent, transit fare
  prices.json             groceries and utilities per place
  wishlist.json           iPhone, Switch 2, Stanley, Starbucks, Big Mac, Netflix per country
  regions-id.json         2026 UMP/UMK for Indonesian regions
  timeseries-jakarta.json Jakarta UMP 2000–2026 and published rice-price anchors
  sources.json            every source: title, publisher, URL, access date
  notes-id.json           Indonesian version of every English note (shown when the page is in Indonesian)
  method.json             basket quantities, workday length, index base year
  wid-indonesia.json      WID income distribution, Indonesia 2024 (127 percentile groups) + conversion factors
  wid-world.json          WID income distribution, world at PPP, 2024
  bps-wages.json          BPS average take-home pay (Sakernas), national / sex / sector
  freshness.json          how often each indicator is published (drives /status.html)
scripts/check-data.ts     validator (npm run check-data)
scripts/og.ts             share images and share pages (npm run og)
src/main.ts               page shell, events, source dialog, URL hash
src/sections.ts           one renderer per section (HTML strings)
src/map.ts                bubble map (d3-geo + world-atlas 110m, loaded on demand)
src/compare.ts            two cities side by side
src/rank.ts               "Kamu di persentil berapa?" (income percentile; inputs never stored or sent)
src/status.ts             /status.html: is every number current?
src/lib/percentile.ts     percentile lookup and conversions (METHOD.md)
src/lib/freshness.ts      freshness rules shared by /status.html and check-data
src/art/                  hand-drawn SVG flags and city landmarks
src/i18n.ts               Indonesian / English strings
src/store.ts              data + reader's edits (localStorage) + UI state
src/data/types.ts         TypeScript shapes of the data files
src/data/units.ts         unit parsing shared by the model and the validator
src/lib/model.ts          the method: pay per minute, work time, basket, interpolation
DATA_UPDATE.md            monthly update checklist
METHOD.md                 how the income percentile is calculated, and the freshness rules

Data rules

Every number is a Datum:

{ "value": 5729876, "unit": "IDR/month", "asOf": "2026-01-01", "sourceId": "kemnaker-ump-2026", "status": "official", "note": "optional" }

field	meaning
value	a number, or null for "not sold here" (a note saying why is then required)
unit	<numerator>/<denominator>: IDR/kg, THB/day, KRW/hour, IDR/USD, h/month
asOf	YYYY-MM-DD the value applies to: effective date, survey date, or the day a shop page was read
sourceId	an id in sources.json
status	official (government or seller), snapshot (observed once: shop page, survey, third-party index), estimate (our approximation), interpolated

• A value that has not been found yet is left out of the file, not set to null and not guessed. The validator lists these as gaps.
• Wages are stored in the unit the law uses (per month, day or hour). The model converts them using monthlyHours and method.workdayHours.
• Wishlist prices may be in local currency or USD (Cambodia prices in USD); the model converts with fxPerUsd.
• Time series store published anchors only. The site draws straight lines between them and marks those years interpolated. The validator rejects hand-entered interpolated points.
• sources.json accessed: null means the figure was carried over from the prototype and its source has not been re-opened yet.

npm run check-data fails on: a value without unit, date, source or status; an unknown source id; a unit that doesn't match the field (e.g. JPY/item for Thailand, a wage per week); an asOf after the file's updated date; null without a note. It prints gaps (missing values, sources not re-opened, values still citing the prototype) as warnings. npm run check-data:strict turns gaps into failures; use it before calling the data complete. Add --verbose to list every missing value.

Method

• Pay per minute = monthly wage ÷ (monthly hours × 60). Work time = price ÷ pay per minute.
• Monthly hours: 173 (ID, JP), 174 (CN), 191 (SG), 195 (MY), 208 (TH, VN, PH, KH, LA), 209 (KR): each country's statutory normal week × 52 ÷ 12 (Korea's official 209), stored per place with the law in note.
• Daily wages count monthly hours ÷ 8 days (e.g. Bangkok ฿400 × 26).
• Monthly basics basket for one person (method.json): one room, 8 kg rice, 3 kg eggs, 4 kg chicken, 1 kg beef, 2 L oil, 90 kWh electricity, 44 transit rides. Leftover = wage − basket.
• Every computed figure carries the weakest status of its inputs, so anything built on an estimate is shown as an estimate.
• Jakarta history uses Bank Indonesia PIHPS (mid-January, DKI Jakarta traditional markets), which starts in 2018. Earlier years are not guessed; published pre-2018 rice anchors can be added to timeseries-jakarta.json and the chart will draw them, with years in between marked interpolated.
• Myanmar is left out: its exchange rate and price data are unreliable.

Design notes

• Indonesia in rupiah red, every other city in slate. Hatched bars, dashed bubbles and a ≈ mark mean the number includes an estimate; the legend says so under every chart.
• Every number on screen is a button: one click opens where it comes from (value, unit, date, status, source link, note), including all inputs of a computed figure.
• Each city carries its flag and a line-drawn landmark (Monas, Petronas Towers, Marina Bay Sands, Wat Arun, Bitexco, a jeepney, Independence Monument, Patuxai, N Seoul Tower, Tokyo Tower, a KRL train for Bekasi, Tugu for Yogyakarta).
• Basket colours (room, food, transit, electricity) are validated for colour-blind separation and 3:1 contrast in both themes.
• Sections render as they near the screen; the map and world atlas load after first paint.
• Bahasa Indonesia by default; ?lang=en (or the ID | EN switch) gives English, and the switch updates the address so it can be shared as-is.
• Cities: Jakarta, Kota Bekasi, Bandung, Yogyakarta, Surabaya, Denpasar, Medan and Makassar in Indonesia (each on its own 2026 UMK and provincial PIHPS prices), plus Kuala Lumpur, Singapore, Bangkok, Ho Chi Minh City, Manila, Phnom Penh, Vientiane, Shanghai, Seoul and Tokyo.
• Bandingkan (compare): pick any two cities; every basic, want and experience is shown in each city's own work time, with the ratio. Shareable as #compare-jakarta_vs_tokyo.
• Wishlist includes experiences: the cheapest official BTS ARIRANG World Tour ticket in each city on the tour (cities without a show say so), and a trip to Europe = return economy flight to Paris + Schengen visa. A bundle item ("bundle": [...] in wishlist.json) has no prices of its own; the site adds up its parts and keeps every part's source.
• URL hash keeps the view: #wishlist-iphone, #item-beef, #map-bigmac, #history-eggs, #regions-umk, #compare-jakarta_vs_tokyo, #rank-t42 (a shared percentile; never the income unless the reader opts in).
• Kamu di persentil berapa? Household income → per adult → position among adults in Indonesia and the world (WID pre-tax national income, 2024), with thresholds, multipliers, work time and growth since 1980, plus a wage lens that compares take-home pay with BPS averages (BPS publishes no wage percentiles). Data credit: World Inequality Database (WID.world). Method and the difference from WID's own comparator: METHOD.md.
• Data status (/status.html): every number's date against its indicator's publishing schedule (): current, due soon or overdue. applies the same rules.

Updating data

See DATA_UPDATE.md.

License

All rights reserved; see LICENSE. Readers may quote individual figures with credit and a link; the code and the compiled dataset may not be copied, scraped, modified or used commercially without permission. The underlying statistics belong to their publishers (named in every source panel) and remain available from them. The site shows the same terms at /lisensi.html, and public/robots.txt asks AI-training crawlers not to collect the site.