Two official tables put Seoul’s population at roughly 9.3 million around the end of October 2025. The figures look close enough to use interchangeably in a caption. However, one comes from the resident register, while the other describes the population usually living in the city. This is where Korea population data becomes a question of definitions.

The distinction matters in everyday reporting. A map of registered residents, a story about households and a forecast of future school-age populations may all start with a population table. Each one therefore needs a different description of who is being counted. The 2025 census gives us a chance to see how those choices work in practice.

Korea population data starts with a reporting question

Imagine opening a spreadsheet for a Seoul neighborhood story. A producer asks for the local population and wants a figure beside the map. First, decide what the map represents. A portrait of the people who usually live there calls for a different table from a chart of monthly changes in the resident register.

Time matters, too. For example, a story about people living in Seoul in late 2025 needs a reference date in that period. A study of future demand for a service needs an estimate for its planning year. Even if the titles use the same word for population, the rows carry different kinds of evidence.

Think of each series as a record of a particular relationship to place. Resident registration records a person’s administrative relationship to a registered address. The census applies a usual-residence concept. Meanwhile, the national projection describes how an estimated population would evolve under specified assumptions.

Before looking up a number, write one sentence about the result you need. For example, the share of general households made up of one person asks for household counts. The number of registered residents per district asks for people on the register. Those sentences are already close to the labels a reader should see.

That small step also makes it easier to choose the right precision. A national overview may need a rounded figure in millions. A reproducible comparison may need the original cells and units. Start from the table, then decide how to display it.

Resident registration counts a monthly administrative population

The Ministry of the Interior and Safety publishes resident registration figures by administrative area. Its monthly table uses the final day of each month as the reference. Therefore, a number published in November can describe October’s month-end population. Keep the publication date and the count’s reference date in separate fields. MOIS monthly population statistics

On the Korean portal, the registration category marked 전체 combines three groups. It includes registered residents with a known place of residence, people registered as having an unknown residence and registered overseas Korean nationals in the portal’s category. The national total excludes foreign nationals. This gives the word total a specific administrative meaning.

For example, open the monthly population and registered-household table, select the national view and choose October 2025 as both ends of the period. Keep the registration category set to all. The national and metropolitan rows then appear together. The page also presents registered household counts, called 세대, beside the person counts.

When comparing two months, keep the same registration category and check the geographic row. A national row contains a national count. Selecting Seoul in the area menu also brings up its districts. Therefore, record whether the figures came from the metropolitan row or a district view.

Monthly frequency makes this table useful for following the register. A reporter could, for example, show how Seoul’s registered population changed over a year. The change can be described directly as a movement in the count. A story about the reasons behind it would need separate evidence about the components of that change.

A register-based census builds a population from several records

The Ministry of Data and Statistics released the 2025 register-based Population and Housing Census results on July 28, 2026. Its population counts refer to midnight at the start of November 1, 2025. The same date also applies to the household and housing results. 2025 census release

Here, registration is a way to produce statistics from several sources. The 2025 results linked 27 types of administrative data, including resident and foreigner registers, building records and university records. The usual-residence criterion covers Korean and foreign nationals residing in the country for at least three months. Official census briefing

The census uses two methods for different sets of questions. The full-count component has used administrative records annually since the 2015 census. The separate five-year sample survey collects more detailed information, including topics such as commuting and education. Specifically, its 2025 fieldwork used a 20 percent sample. English census survey guide

That division matters when opening a historical chart. The Korean report labels selected older field counts with a T and register-based counts with an R. A change over five years and a change over one year also refer to different intervals. Accordingly, keep the method marker and interval alongside any values transcribed from a long series. Full Korean census report

For a 2025 population count, the full-count release is the starting point. For a commuting question, first find out whether the requested detail belongs to the sample series. Check the table notes because a census title can refer to either component.

Compare population counts around the same date

Start with the national rows. At October 31, 2025, the resident registration total was 51,143,421. The census reported 51,817 thousand people at midnight on November 1. Those are the closest published reference labels used here, and each number describes its own population. October 2025 resident register table

The register displays its count in single persons. By contrast, the census release rounds the national total and its nationality categories to thousands. For example, it separately reports 2,109 thousand foreign nationals in the census population. That category has its own row and usual-residence definition. Use that row when describing foreign nationals in the census. English census tables

Foreign-national coverage is one reason the totals answer different questions. Conversely, the official census briefing explains that the resident register can include Korean nationals living overseas for extended periods. Both coverage rules matter before making a comparison. Use a table that measures the group directly because the difference between the totals combines several effects.

Putting the rows side by side still serves a purpose. It shows how much context a short label like South Korea’s population leaves out. A chart reader should be able to see the two reference dates, the different definitions and the display unit without opening either source.

Finally, keep both values tied to their own publication. An October register number belongs in a chart about registration. A census number belongs in a chart about the usual-resident population as defined in the census. This gives the comparison a purpose beyond choosing the larger or more recent figure.

Seoul and Incheon show how population counts vary by area

In Seoul, the October 2025 resident registration count was 9,313,532. The census report’s provincial table gives Seoul’s November 1 population as 9,315 thousand. Both become about 9.3 million when displayed to one decimal place in millions, although their sources use different inclusion rules. Census report, Table 5

Next, look at Incheon in the same tables. The register gives 3,049,315 people for October. The census reports 3,094 thousand for November 1. In a table rounded to thousands, the two Incheon figures sit farther apart than the two Seoul figures.

The two examples show why the scope belongs in every regional table. Similar totals in Seoul can hide distinct definitions. Meanwhile, wider separation in the Incheon rows gives no breakdown of the people behind it. A separate nationality or residence table is needed to describe those groups.

Then there is a third number a reader can encounter. Seoul Metropolitan Government’s English city profile shows 9,560,471 people for the second quarter of 2026 and identifies 270,658 foreign residents in its figures. Its quarterly reference and its inclusion of foreign residents distinguish the city profile from the October 2025 resident register row. Seoul city population profile

Save the area label as well as the number. Seoul means the city in these rows. A figure for the wider capital region includes other places. Also, a national ranking by cities needs a consistent definition across the rows, even if a city’s own profile is easier to find.

Match population counts to the boundaries of the story

Place names have a way of expanding in conversation. For example, a person may describe an address outside Seoul city as being in the Seoul area. A national statistical table can distinguish Seoul from the wider capital region, which also covers Incheon and Gyeonggi. The chosen geographic label determines the population being described.

A district-level project needs another decision. A district boundary may fit a story about its resident population. Meanwhile, a story about a train station’s catchment might draw on several districts or on travel data. The number of people who live within a boundary and the number who use a place during the day measure separate things.

Consider a hypothetical comparison between two neighborhoods with a similar number of residents. One has a busy transfer station and the other is mostly residential. An address-based population count can help describe where people live. However, counting visits to a station calls for transport or footfall data matched to the time being studied.

Seoulz’s reporting on rental markets offers a related example of a geographic distinction. Its coverage of monthly rent presents district figures and separates apartments from other property types. A reader comparing rent patterns needs the same type of property as well as the same geographic scale. Seoulz coverage of monthly rent by district

For a time series, also check whether the boundary or area code changed. When the historical table uses a different geography, a comparison may require a consistent boundary series. Save the printed names and codes so another reader can see how each row was selected.

Pick a unit before comparing people and households

Population counts can sit next to household and housing counts on the same page. First, read the unit in the column header. The census measures people, households and housing units separately. For example, the household figure describes living arrangements. The housing unit figure measures dwellings.

The Korean census report also distinguishes a census household, 가구, from a registered household, 세대. Its appendix describes a census household as a living unit of one or more people who share daily living arrangements. The resident register groups people for administrative purposes. Therefore, even the English word household needs a source label. Census concepts and comparisons, appendix

Within the census, general households are a category with its own total. The report also presents a broader total that includes group and foreign-national households in separate categories. The classification is about households. When a table breaks down individual people by nationality, it uses a different unit and set of categories.

This matters when reading about solo living. Seoulz’s earlier coverage of the solo economy compares widely cited figures and explores what businesses sell to people who live alone. Here, the task is to recreate one census share. That requires identifying the specific category of households behind it. Seoulz coverage of Korea’s solo economy

If you need an address count, a measure of occupied dwellings or a number of potential customers, define that measure on its own terms. A household count can help describe a market’s context. It still leaves questions about the people, properties or purchases that make up the market.

Build a household share from the same census table

A practical use of Korea population data starts with a real calculation. In the English 2025 census report, find the table of major indicators and go to the 2025 column. Its household section shows 8,244 thousand one-person households and 22,495 thousand general households. Census major indicators

First, take the one-person household row as the numerator. Next, take the general household row as the denominator. Dividing 8,244 by 22,495 and multiplying by 100 gives about 36.6 percent. The thousands cancel because both rows use the same unit. That result reproduces the share reported in the release.

The denominator gives the result its meaning. Here, one-person households are a subset of the 22,495 thousand general households. If you use the report’s broader total of about 23.25 million households, you answer a different question. A population total would change the unit from households to people.

The procedure also works for regional shares if both rows come from the same regional breakdown. For example, the release reports that 40.5 percent of Seoul’s general households were one-person households. The nationwide result answers the same type of question at a different geographic scale.

Finally, keep rounded numbers rounded. The report’s numbers are in thousands, so the calculation reproduces the published percentage at one decimal place. Write the result as a share of general households on November 1, 2025. That phrasing gives readers the group, place and date they need.

Population projections start from a stated base

The national 2022–2072 projection was released in December 2023. Its starting population was about 51.67 million on July 1, 2022. The release used the 2022 census and demographic information to estimate that midyear base. A projection chart therefore starts from a defined date and a particular set of source data. National population projection release

The projection then follows age and sex groups over time. In broad terms, it ages the starting groups, uses mortality assumptions to estimate survival and accounts for births and international migration. This is called the cohort-component method. It lets the totals and the age structure change together.

For example, the medium scenario puts South Korea’s population at about 36.22 million in 2072. The year and scenario belong next to that figure. The base value describes an estimate for 2022. By contrast, the 2072 figure describes the population under the scenario assumptions. Projection methods and scenarios

A line drawn between those values can look much like a history chart. In a graphic, consider showing the estimate and projection with separate line styles. Also, if you add a census observation to the chart, label its November reference date. The projection series uses a July 1 population reference.

When reporting a target year close to the present, check how the data series labels that year. A projection published before a census can still contain a value for the census year. The two numbers should retain their own reference dates, methods and labels.

Read population projections with the scenario attached

In the same 2022–2072 release, the high and low growth scenarios put the 2072 population at about 42.82 million and 30.17 million, respectively. Together they illustrate how the modeled outcome changes when the assumptions change. A reader who sees all three figures can separate the medium scenario from the alternatives.

Those assumptions have a direct role in the calculation. For example, changes in fertility influence the number of births in the model. Mortality assumptions shape how many people are expected to survive to later ages. International migration can change the size and age mix of the population under a scenario.

Use a scenario to test a planning question. For example, an organization exploring future demand for a service can see how a different population path might affect it. It would also need assumptions about participation, where potential users live and how the service operates. Therefore, the population projection supplies one input to the planning exercise.

A national scenario also has a geographic limit. The Ministry of Data and Statistics lists a separate provincial projection with a 2022–2052 horizon. For a Seoul planning question, check the release that actually contains Seoul’s geography, time span and age groups. Official projection release catalogue

Finally, save the release date with the value. A later projection can start from different evidence and use revised assumptions for the same target year. Keeping a record of the release makes it possible to explain which version appeared in the original chart.

Choose population counts that match the numerator

The household example is a special case of a broader rule. Define the people or units in the numerator, then choose the denominator that describes the group from which they are drawn. A good denominator follows the geography, time, unit and inclusion rules of the measure.

Suppose a hypothetical city program says 600 Seoul residents used a service during a quarter. First, check what the program calls a resident. Its own eligibility rule might refer to registered address or include people who demonstrate residence in other ways. The choice of denominator therefore follows the program’s definition.

Next, check whether the numerator is people or visits. If one person can appear more than once, dividing the visit count by the resident population gives visits per resident. A share of residents who used the service requires a count of distinct people. The two measures answer different questions about reach and activity.

The period also needs attention. A monthly population figure is a snapshot. A quarterly service figure accumulates across months. For example, an agency might use a stated midperiod population for an event rate. Record that denominator convention and apply it consistently across the areas or periods being compared.

Finally, check age and nationality whenever the numerator selects a subgroup. A count of residents aged 65 and over calls for a denominator whose age groups and population coverage support the intended question. In particular, the label should tell readers whether the result is a share of all residents or a rate within that age group.

Leave a record someone else can follow

A compact record of Korea population data can fit beside each number in a working spreadsheet. Start with the name of the agency and table. Then save the link, the reference date and the date you retrieved it. Keep the geographic label, population definition and unit with the original value.

For the Seoul comparison, that means two rows. One says Seoul’s October 2025 month-end resident registration total and preserves the exact count. The other says Seoul’s November 1, 2025 register-based census population and preserves the figure in thousands. The source descriptions travel with the cells.

For a percentage, record the numerator and denominator as separate cells. Also save a short formula and the published precision. The 36.6 percent household example can be reproduced from the major indicators table because the underlying household rows are visible in the working record.

For a projection, add its release date, base and scenario. A future number can then be distinguished from an observation when the spreadsheet is reused. If a newer release is adopted, both the revised number and the reason for updating the chart can be traced.

The last step is editorial. Read the caption without the spreadsheet open and see whether it still says who was counted, where and when. A caption that passes that test gives a reader enough information to choose between the two Seoul numbers and understand what each one can say.