🗺️ GeoWho

A daily "guess the historical figure" game. A spinnable globe marks where a famous person was born (★) and where they died (✝), with their exact birth and death dates. Drag to rotate, scroll/buttons to zoom, then name who it is.

Three figures a day, with a scoring ladder per figure:

When you get it	Points
First try, from the map + dates alone	100
After one free hint	50
From four multiple-choice options (after a second hint)	10
Not at all	0

Each wrong guess automatically reveals the next hint (category → occupation → country). After the third figure you get your total score out of 300, a shareable spoiler-free result, and a countdown to the next day's three. Progress is saved, so the day resumes where you left off.

There's also a Practice mode: unlimited random figures with the same scoring ladder and a running session tally — no daily limit.

Run it

npm install
npm run dev        # http://localhost:5173

Other commands

npm run build      # type-check + production build to dist/
npm run preview    # serve the production build
npm run verify     # data/logic tests + full render smoke test
npm run data       # rebuild the dataset from Wikidata (see below)

How it works

• The day's three figures are chosen deterministically from the date, so everyone gets the same puzzle; no repeats within a day.
• Guess with an autocomplete name box (ranked most-famous-first); the final stage offers four multiple-choice options — same category, closest era.
• The dataset is trimmed to the top 200 most famous figures (household names), with real birth/death coordinates and dates. TOP_N in scripts/prep-people.mjs controls the size.

Code layout

• src/lib/game.ts — the game rules: daily selection, random pick (practice), the pure scoring reducer (applyGuess/applyPick), option builder, share text, date helpers, constants.
• src/lib/people.ts — dataset access, search, clues.
• src/lib/storage.ts — daily progress persistence (localStorage).
• src/App.tsx — mode switcher (Daily / Practice).
• src/components/ — Daily and Practice (the two mode orchestrators), Round (the per-figure stage machine), WorldMap, GuessInput, Options, Clues, Summary.

The data (Wikidata)

A global "most famous humans" query times out, so the pipeline scopes by occupation (small sets) and takes the top figures by fame within each:

Script	Role
scripts/fetch-people.mjs	Top figures across ~35 occupations, with birth/death coords + dates.
scripts/fetch-people-extra.mjs	Big/late occupations needing a higher fame threshold (politician, statesperson, naturalist).
scripts/fetch-people-must.mjs	Guarantees curated iconic figures by QID.
scripts/fetch-dates.mjs	Adds each date's precision so exact dates are honest (no fake "1 January" for year-only records).
scripts/fetch-occupations.mjs	Adds real P106 occupations, used to categorize each figure (clue #1).
scripts/prep-people.mjs	Formats dates and emits src/data/people.json with each figure's raw lng/lat (blng/blat/dlng/dlat).

WorldMap.tsx renders a realistic orthographic globe — filled continents, country borders, and a graticule — with d3-geo + topojson-client over the world-atlas land geometry (loaded at runtime). Drag rotates, wheel/buttons zoom, and the globe auto-faces the midpoint of the two locations.

Dates render at their true precision: day (14 March 1879), month, or year (550 BC). A few otherwise-iconic figures can't be included when Wikidata has no coordinates for their birth/death places (e.g. Frida Kahlo).

Deploy

Fully static — npm run build and host dist/ anywhere. This repo also auto-deploys to GitHub Pages on every push to main (.github/workflows/deploy.yml).