Staffed hours can vary
Some shops switch between staffed and unstaffed hours. Meal-kit chain Damgguk recommends staff during the day and self-service at night.
Damgguk’s operating example ↗THE SELF-SERVICE CITY 무인점포
Seoul, Unmanned.
Explore Seoul’s self-service shops, from laundromats and study cafés to ice-cream stores.
4,251 surveyed records. 25 districts. Six kinds of self-service.
2025 FILE SNAPSHOTWhat are these shops? ↓
We couldn’t load the map. Please reload or download the survey summary.
SEOUL / ALL TYPESSELECT A DISTRICT
Circle area = recorded count
District totals, not shop locations.
A surveyed subset, not a census. Coverage is unknown; a small count doesn’t mean few shops. About the data ↗
BEHIND THE COUNTER
How the shops operate
Someone can still be watching
All 30 unmanned ice-cream shops in a 2023 consumer-agency inspection had CCTV. This was a small sample across three regions, including Seoul.
Korea Consumer Agency, 2024 ↗Are more shops doing better? NATIONAL DATA +
In Samsung Card’s nationwide sample, active unmanned merchants grew 19% from June 2023 to June 2026. Yet six of eight categories had lower average sales per shop in June 2026 than June 2024. More shops does not necessarily mean better business.
A different sample from the Seoul survey. Sales are not profit; these figures alone do not establish saturation. Samsung Card, August 2026 ↗
Sources and limitations DATA NOTES +
What you’re looking at. 4,251 records from Seoul’s Fire and Disaster Headquarters, covering six types of unmanned business across 25 districts. This is a surveyed subset, not a complete business census, and not 4,251 verified unique shops. Hybrid operations and actual opening hours are not specified.
Which date? The file is named 무인점포 현황_20251231.csv. We label it a 2025-file snapshot; the exact fieldwork dates are not supplied. The downloadable catalog was published April 6, 2026; its attachment was modified April 29. A September 30, 2026 catalog lists the same 4,251 count but no file. We cannot establish that it contains a newer survey or identical records.
What the map means. One circle per district, with area proportional to its recorded count. The scale stays fixed when you change category. These are raw counts, not population-adjusted rates or a ranking of retail density. The exact counts accompany each share. Guro has only 19 records, Yangcheon 46, and Eunpyeong has no study cafés recorded. None of these establishes how many shops actually exist there. “Gu” means district.
Repeated names. There are 57 repeated district/name/category combinations, producing 64 excess rows. Without addresses, some could be different branches using the same name. We retain all rows. Removing exact repeated combinations changes each featured district share by less than two percentage points. Four more combinations merge if whitespace is trimmed. Three district/name pairs appear in two categories, so the categories are not guaranteed to represent distinct physical outlets.
No invented locations. The source has only row number, district, business name and category. No addresses, coordinates, opening dates, staffing schedules or closure dates. Circles are district aggregates, not shop pins. Simplified district outlines come from KOSTAT 2013 via southkorea/seoul-maps (attribution, Apache 2.0). They provide orientation, not current legal boundaries.
Other research. Samsung Card, August 2026, uses nationwide card merchant and transaction data. Active-store growth compares June 2023 to June 2026; sales per shop compare June 2024 to June 2026. This is a different sample with eight categories. Sales are not profit, and the report does not establish a same-store or causal saturation effect. The KCA inspection covered 10 shops each in Seoul, Gyeonggi and Chungcheong; it is not a representative CCTV prevalence estimate for Seoul.
Historical counts. Earlier police and fire counts use different coverage and definitions. We found no verified comparable series for the supplied Seoul file. The dated moments are separate examples, not points on a growth chart.
Download the calculated summary ↓Sources checked October 17, 2026 · No tracking · No live data calls