Topics that shape the future
- Closer look at research results -

It can also be used in school consolidation plans. Because it allows for predictions of future student populations in detailed areas, it becomes important data for considering how to place schools. In addition, it can be used in various regional policies such as countermeasures against tsunamis and floods, and the formulation of disaster prevention plans. I personally feel that future population projections for small areas have virtually limitless potential applications.

As mentioned above, I began working on future population projections for small regions after taking on a project for the Tokyo Metropolitan Government, so that was about 12 or 13 years ago. Before that, I mainly focused on regional population theory, particularly the distribution of population and migration between regions. My focus on regions was largely influenced by the fact that my research foundation was in geography. As I mentioned in a previous AGU RESEARCH presentation, I have loved looking at demographic data and maps since childhood, and I was engrossed in them as a boy. On the other hand, I also developed a strong interest in mathematics in junior and senior high school, and I went on to study at the University of Tsukuba, School of Natural Sciences (at the time), where I could specialize in either mathematics or geography. During my undergraduate years, I diligently studied probability theory and statistics, but I began to feel the limitations of pure mathematics, so I decided to switch to population geography, which had been my interest since childhood, for graduate school.

Population geography is a field within human geography that uses a relatively large number of mathematical formulas. Therefore, I found a way to utilize what I had learned and apply mathematics. Pure mathematics is not merely calculation; it is a field of study that includes philosophical elements, a world that cannot be fully explored without innate talent. In that respect, I thought that the field of population geography would offer an opportunity to demonstrate the mathematical thinking skills I had acquired.

At the University of Tsukuba, I had the opportunity to study not only mathematics but also geography from a scientific perspective. Thanks to this, I was able to arrive at the field of study that combines demography and mathematics, which I had been interested in since childhood, and I am still engaged in research there today. Looking back, I feel that all of my experiences have had a positive impact. Of course, the fact that I was able to successfully change my specialization is thanks to the guidance of the professors in the geography department during my graduate school days, especially my mentor, and I cannot thank them enough.

This award was established in 2005, and I actually shared the first award with others for my work on "GIS education in university College of Economics." This is my second award, 19 years later. This award is given to individuals or organizations that have achieved outstanding results in activities related to "statistical GIS," which integrates geographic information systems (GIS) and statistical information. I believe that my series of studies on future population projections for small areas were recognized as fitting the purpose of this award.

As I mentioned earlier, the impetus for working on "future population projections by small area" came from a request for advice from the Tokyo Metropolitan Government regarding future population projections by small area. What I realized at that time was that there were hardly any rational methods developed for population projections at the small area level. I think that few people were working in this field because the projection results tend to be unstable when the regional unit is small, and the necessary population data was not sufficiently compiled.

The biggest challenge in estimations for small regions is that the estimated values fluctuate wildly. The methods considered so far have generally involved a simple process of truncating the fluctuations at a certain value. However, this treats extreme fluctuations and almost no fluctuations the same, resulting in inaccurate results. Therefore, I devised a method to smooth the parameters necessary for estimation and introduced a mathematical formula that enables more precise estimations.

Smoothing is a process used to suppress anomalous values in parameters required for future population projections that could lead to unusual results. This smoothing method utilizes the concept of population potential, a classical theory in geography. By formulating this mathematically and combining it with existing future population projection methods (such as the Hamilton-Perry method, known in Japan as the cohort change rate method), the projection results become more stable.

Population potential is defined by the principle that "the demographic influence of region B on region A is proportional to the population of region B and inversely proportional to the distance between regions A and B," thus defining the sum of the influences of surrounding regions on region A. This concept is a fundamental idea in geography, but it had never been applied to future population projections until now. Focusing on this point ultimately led to the development of a new projection method.

The resulting projection system was published online as the "National Small-Regional Population Projection System" (SAPP for Japan) and is used by many researchers and practitioners in local governments and other organizations. Furthermore, the research itself has been cited by researchers worldwide, and we have received invitations to speak at international academic conferences. In addition, based on these findings, we have built systems targeting the United States, Taiwan, South Korea, Australia, and other countries, and have published them as the "SAPP Series."