Use Seoul neighborhood population estimates to plan local ads

Where should a lunchtime promotion run?

Choosing a Seoul neighborhood for a lunchtime promotion takes more than its resident population. Commuters and shoppers can change the area’s population throughout the day.

In Neuro, you can query Seoul’s estimated living population by administrative neighborhood, or dong, date, hour, and age group. These estimates do not measure store visits or purchases, but they can help narrow down places and times to compare.

What the data looks like inside Neuro

The dataset we connected here is the administrative-dong–level Seoul Living Population (Korean nationals) from the Seoul Open Data Plaza. Per the official description, it estimates the population present in a given area at a given time by combining Seoul city public data with telecom data. In Neuro, we structured it so you can query by date, daypart, administrative-dong code, city/gu and dong name, total living population, and living population broken out by gender and age group.

When you name a region in natural language, Neuro first searches candidate administrative dongs to lock in the exact one, then queries the living population table. A request like “Show me the weekend-evening living population for 20-somethings in Seongsu-dong,” for instance, has to be split into place-name resolution and time conditions handled separately. That’s how we cut the risk of reading the wrong area because of similar dong names or vague district-level phrasing.

Putting it to work in franchise F&B

Consider two hypothetical coffee shops. You might first examine weekday lunchtimes in an office district and weekend afternoons in a residential area. These are starting hypotheses. Check whether the hourly neighborhood estimates support them, then compare store sales and coupon use.

• In areas with a sharp lunch peak, lead with messaging around quick table turnover, set menus, and mobile ordering.
• In areas where living population climbs after work, look at delivery, takeout, and dinner promotions together.
• In trade areas with big weekend swings, don’t cut budget off weekday-average ROAS alone — re-split it by day of week.
• Before opening a new franchise location, compare daypart living population across candidate dongs against the sales patterns of existing stores.

The point is not to use living population as a targeting option. Ad-platform targeting and living population are different datasets. In Neuro, living population is an input for “what regional and daypart hypotheses should we form,” while the actual performance read has to come from ad metrics, store sales, coupon redemptions, and order data together.

Putting it to work in OOH and local media

OOH is bound to place and time even more tightly than online advertising. When you evaluate subway, bus, billboard, or elevator media, “lots of foot traffic” isn’t enough to go on. You need to see which dongs run strong by day versus by night, how wide the weekday-versus-weekend gap is, and which dayparts make your brand’s target age groups stand out the most.

With this data in Neuro, you don’t jump straight to a buy plan on broad units like “all of Gangnam-gu” when shortlisting OOH placements — you can narrow candidates down to the administrative dong and daypart. The bigger the media budget, the more you should commit a hypothesis to writing before you buy. Logging why you picked this area, which population signals you saw at which dayparts, and how you expect it to split roles with your online campaigns is what lets you sharpen the next placement.

How to ask Neuro

Confirm the administrative neighborhood first. A dong estimate is not the population within a 500-meter radius of a store. Check the latest available date and compare matching weekdays and hours.

1. Confirm the administrative-dong name first. If several dongs share a name or you’ve only given the district, Neuro surfaces candidates and asks you to confirm.
2. Set the period and dayparts. Decide whether to look at daily averages, split weekday from weekend, or align with a specific campaign window.
3. Read total living population alongside the gender and age breakdown. Even when the overall volume is high, you need a different play if your target segment is thin.
4. Tie it to other data — ad performance, store sales, coupon redemptions, search trends. Don’t lock in budget decisions on living population alone.
5. Finally, write down your execution hypothesis. Logging which messages you’ll test in which regions and dayparts is what makes the next analysis easier.

What not to read into it

Living population is an estimate. The official documentation itself notes that it can diverge from reality for certain areas or moments in time. The public data doesn’t track individuals, and it’s not a real-time sensor telling you how many people are standing in front of your store right now. So don’t read it as purchase intent, probability of a visit, or guaranteed ad impressions.

Drawing conclusions about an area from a single day of data is especially risky. Weather, events, public holidays, school calendars, large rallies, and concerts all creep in easily. When you use living population in Neuro, at a minimum set comparison periods, compare like-for-like days of the week, and read it alongside campaign performance. More data doesn’t automatically make a decision more accurate.