02 Core Concept: How the AI works
The system scores each candidate shelter across six factors:
Recommendations are delivered in plain language via RAG (Retrieval-Augmented Generation) — the AI only references verified public data, never inventing facility details.
/ 01
Distance
Walking distance from where you stand
/ 02
Availability
Open now, with a number to call
/ 05
Safety
Avoids senior pedestrian-accident zones
/ 06
Data reliability
How recent and verified the record is
/ 03
Weather fit
Designated heat- or cold-shelter status
/ 04
Transit access
How close the nearest bus stop is
03 Public data sources
Gyeonggi Data Dream
-senior welfare facilities
-senior classrooms
-heatwave shelters
-warmth shelters
-bus stops
-senior pedestrian accident hotspots
KMA APIs
-weather forecasts
-alerts
-impact forecasts.
04 Key UX principle
Designed for elderly users
-goal reachable in 4 taps or fewer
-with large text
-voice search
-one-touch phone call to the shelter, 119, or a caregiver.
05 Business model
B2G pilot with Yongin and Seongnam cities (~₩20M/city)
scaling to all 31 Gyeonggi municipalities by 2028, then nationwide.
BUSINESS PLAN FULL VERSION TBA
01 Problem Definition
Map apps point to the nearest pin.
For an older resident in a heatwave, a shelter 250 m away can be the wrong choice — closed, with no air conditioning, or across a pedestrian-accident hotspot.
A shelter 430 m away may be the one that actually keeps them safe.