Yukikaki Link (ゆきかきリンク)

Connects the neighbor who needs snow shoveled with the neighbor who can help — three taps, no accounts, no money, works offline. Built for Sapporo, where 29% are 65+ and snow is a daily risk.

Comment

What is the specific problem, and how do you know it's real?

In Sapporo, Japan — my city — winter snow is not just a chore. For many people here, it's a daily physical risk they face alone.

I know it's real because I've seen it on my own street: on a December morning, our elderly neighbor was bent over, clearing the snow in front of her door by herself. When we asked why she didn't get help, she said: "I have no one to ask." That sentence is where this project started.

And the public data confirms it's a systemic gap, not a one-off:

  • 566,710 residents of Sapporo are 65 or older — 29.35% of the city (Hokkaido residency records, Jan 2026). 121,789 senior households are single-person households (2020 Census).
  • The city's welfare snow-removal program requires care-need certification and that there be no children living within ~500m — so it misses the "still healthy, but shoveling is genuinely hard" households. It reached only 5,692 households in FY2024, and the city itself names "increasing users and a shortage of local helper volunteers" as the program's problem.
  • 1,869 people were taken to hospital after falling on snowy streets in FY2024; 45.7% were 70 or older (Sapporo Fire Bureau).
  • Last winter in Hokkaido, snow caused 175 casualties including 10 deaths — every death happened during snow-clearing work (Hokkaido, 2025-26 season).
  • The city spends ¥28.4 billion (≈$190M) a year on snow measures — but that's for roads, not for doorsteps. Meanwhile neighborhood-association membership has fallen from 83% (1989) to 68.4% (2026): the traditional mutual-help layer is thinning exactly where it's needed.

Every number is cited with its official source in our public evidence kit (link below). The interpretation: between the government program's strict eligibility and the shrinking neighborhood networks, there's a gap — and meanwhile, there are plenty of people who want to help (students before club, retirees with time) but have no point of contact with the people who need it.

What it does

Yukikaki Link is a tiny web app (PWA) that fills exactly that gap:

  • Ask in 3 taps: ① tap your block on the town map → ② what (entrance / path / parking) → ③ when (now / today / tomorrow morning). That's the whole request.
  • Help: a board lists requests from your town, nearest block first. Tap "I'll help" → mark complete → both sides exchange "thank you"s.
  • A "thank-you economy" instead of money: no payments, no points to buy — gratitude stamps pile up on the town's shared counter, and Mikan, our knitted-hat mascot, grows with every act of help.
  • Built for the people who need it: an elder mode with bigger text and furigana readings, gentle Japanese throughout, and no account required — everything stays on the device (localStorage), so there's no personal data on a server.
  • Safety first: roof-snow removal is not requestable (it's where most fatalities happen) and the app explains why; it links to the city's volunteer-activity insurance.
  • Works offline (service worker): a snowstorm is exactly when connectivity is flaky.

Built with

Vanilla HTML / CSS / JavaScript — zero dependencies, no framework. Canvas (town map + mascot), Service Worker (offline), Web Audio (procedural chime), localStorage (device-local data). Tests: Playwright-driven E2E smoke suite, 35 checks green against both local and the live URL.

AI & open-source disclosure

AI coding assistance was used (AI assistant for code generation and QA) and is disclosed here and in the README. All code was written new, during the hackathon window (Oct 16–) for this repository. No pre-existing projects were reused; no OSS libraries are bundled (zero dependencies — Playwright is a dev-only test tool).

Challenges we ran into

Designing for people who don't use apps (elder mode, gentle Japanese, 3 taps maximum); making a matching service with no accounts (local-first + QR invites for real neighbors); drawing a line on safety (roof snow excluded); and keeping it all in one dependency-free file so it genuinely runs on any phone.

What's next

Pilot with a real neighborhood association in Sapporo; wire in the city's activity insurance and community channels; QR-invite distribution through town associations and schools; and measuring completed requests per neighborhood over one winter as the impact metric.

Built With

Share this project:

Updates

Submission history