“Tourism × Spatial Economics”: Understanding Tourism from Space: New Applications of Data in Economics

1. Various types of data used in spatial economics

I am an economist, specializing in a field known as spatial economics.The characteristics of economic activity vary greatly from region to region. While some areas, like Tokyo, are hubs for a wide range of economic activities, others—such as rural areas facing depopulation—are losing their economic vitality. Spatial economics is the field that examines the distribution and differences in economic activity across regions, as well as the underlying economic mechanisms. My research focuses particularly on data-driven analysis, and in recent years, a variety of new data sources have been increasingly utilized in this field.

For example, smartphone GPS data—which gained attention during the COVID-19 pandemic—allows us to track human movement with a high degree of spatial and temporal granularity. I myself have used this smartphone GPS data to analyze the decline in consumer activity in central Tokyo during the pandemic. In addition, data that provides detailed insights into consumer behavior—such as credit card transaction data and point-of-sale (POS) data from supermarkets and other retailers—has seen increasing use in research within spatial economics and economics as a whole in recent years.

Such new data is extremely useful for policy evaluation because it provides timely information that cannot be obtained from data traditionally used by many researchers, such as government statistics. For instance, during the COVID-19 pandemic, when policy decisions had to be made amid rapidly changing circumstances, it would not have been practical to wait for government statistics—which require time from survey collection to tabulation and publication—to assess the current situation. In such situations, the value of this new, highly timely data became particularly evident.In other words, this new data complements the information that government statistics—which are important for capturing long-term trends and providing reliable, systematic data—cannot capture.

Furthermore, such data is also very useful for analyzing tourism. For example, the smartphone GPS data mentioned earlier can accurately capture human flow, making it possible to identify which locations have high foot traffic.In fact, we are already seeing an increasing number of studies that use smartphone GPS-based human flow data to analyze tourists. Similarly, research has been conducted using credit card transaction data to examine the purchasing behavior of tourists during their stays. As such, this new data—whose use has been expanding in recent years in the fields of spatial economics and economics—is also highly valuable for analyzing tourism.

In fact, there is another type of data that is very commonly used in the field of spatial economics: satellite imagery. Some of you may have seen images of the Earth at night captured by satellites.Although no national borders are drawn on the images, the shapes of countries like Japan and the United States emerge through the light. This is because there is a strong correlation between economic activity and nighttime brightness. In areas with high economic activity, the city glows brightly at night due to artificial lighting from streetlights, shops, and other sources. By utilizing this characteristic, researchers are now able to gauge the scale of regional economic activity based on nighttime brightness and conduct various economic analyses.For example, satellite imagery data on nighttime luminance has revealed a decrease in nighttime light levels during the COVID-19 pandemic, particularly in downtown areas where this decline was significant due to reduced operating hours at restaurants and bars. Such data is important because it provides timely, supplementary information that official statistics cannot capture, even in developed countries with robust government statistics. Furthermore, it is utilized in various studies as essential data for gaining a detailed geographical understanding of economic activity in developing countries where data is scarce.

In recent years, the use of satellite imagery has expanded beyond nighttime light levels to include images captured during the day. Daytime satellite imagery is becoming increasingly high-resolution; currently, the highest-resolution commercially available satellites have a spatial resolution of 0.5 meters.This means that the distance between each pixel in the image corresponds to 0.5 meters; with this spatial resolution, it is possible to identify objects several meters in size, such as automobiles. By combining this with machine learning, it is possible to recognize various objects on the ground, which is currently greatly expanding the possibilities for research.

We are conducting research that utilizes data derived from these daytime satellite images to assess economic activity through human traffic, and applies this to the evaluation of tourism-related policies. Specifically, we use the number of cars traveling on roads to gauge regional economic activity through human traffic, and thereby analyze the effectiveness of tourism-related measures.The number of vehicles traveling on roads can be considered a representation of the flow of people in that area. Of course, in the case of major roads such as highways, vehicles passing through a certain point may simply be transiting the area on their way to other destinations and may not necessarily have business in that specific region. However, vehicles traveling on smaller, more local roads are highly likely to be in the area for some reason related to the surrounding area, and it seems reasonable to assume that they have a close relationship with the local economic activity.From this perspective, the number of cars on the road can be considered a proxy for the economic activity in that area, as reflected through the flow of people.

Using high-resolution satellite imagery and machine learning technologies, as mentioned earlier, it is possible to comprehensively identify vehicles from satellite images, and there are already private companies offering such services. By utilizing these technologies, it becomes possible to gauge the economic activity of a specific area based on the number of vehicles observed there.
 

2. Case Study of Data Utilization on Cebu Island

Following this line of thinking, we examined the economic impact of the opening of the new international terminal at Mactan-Cebu International Airport—the gateway to Cebu Island in the Philippines, a tourist destination very popular among Japanese travelers.

Most tourists visiting Cebu Island use Mactan-Cebu International Airport. During the 2010s, the airport reached full capacity due to the increase in tourists, leading to problems such as congestion and delays.Consequently, an airport expansion project began in 2014, and in June 2018, the new international terminal opened, expanding Mactan-Cebu International Airport’s capacity to approximately three times its previous level. Following the opening, the delay rate improved significantly, particularly for international flights, and the number of international passengers also increased.

So, how much of an impact did the opening of this new international terminal have on Cebu’s economic activity through tourism? Since most tourists visiting Cebu use Mactan-Cebu International Airport, it might seem sufficient to simply count the increase in the number of international passengers at the airport. However, there are several issues with this approach. First, it is difficult to assume that all international passengers using Mactan-Cebu International Airport are tourists.Among the passengers at Mactan-Cebu International Airport, there are likely business travelers, international students (Cebu Island is very famous for language study abroad programs), or simply transit passengers. It is not that easy to isolate only those traveling for tourism purposes from this group. Furthermore, where exactly tourists visit in Cebu is also an important issue.Whether the increase in tourists generates economic benefits across the entire city of Cebu, or only in specific areas—such as resort areas—is a crucial point to consider when evaluating the economic impact of tourism. It is difficult to engage in such a discussion based solely on Mactan-Cebu International Airport passenger data.

To address these issues, it is useful to have data that can determine whether the flow of people presumed to be tourists has increased in the Cebu metropolitan area following the expansion of international flights, and where exactly this increase has occurred. Satellite imagery data is particularly effective for this purpose. By identifying vehicles on roads using satellite imagery and verifying the density of traffic in each area, as well as how this has changed before and after the opening of the new international terminal, it is possible to measure the economic impact.If the opening of the new international terminal has led to an increase in international tourists, the flow of people in areas where these tourists are likely to visit should also have increased, and this can be determined by the number of vehicles. Since there is no railway system in the Cebu metropolitan area, roads are inevitably used for intra-city travel. In fact, it was found that after the terminal opened, the overall density of vehicles on roads across the Cebu metropolitan area increased by about 10%.

So, is this increase in foot traffic—measured by the number of vehicles in Cebu—actually due to the airport expansion? Of course, satellite imagery cannot tell us whether tourists are actually inside those vehicles, but with some ingenuity, we can still provide some indirect evidence. For example, focusing on timing is effective. Tourism has peak and off-peak seasons. In Kyoto, for instance, more tourists than usual visit during the autumn foliage season or the Gion Festival.If we observe an increase in foot traffic in Kyoto during these periods, we can reasonably attribute it to tourists. A similar approach can be applied to this case. Cebu also has peak and off-peak seasons for tourism. Furthermore, these periods differ between international and domestic tourists. Therefore, if this increase in the number of vehicles is observed more frequently during the peak season for international tourists, it can be considered a result of the increase in international tourists.In fact, our research found that the number of vehicles increased more significantly during peak periods for international tourists in Cebu, and that this phenomenon was particularly pronounced on Mactan Island, where resort hotels are concentrated. This suggests that the increase in foot traffic in Cebu City is driven by the rise in international tourists following the opening of the new international terminal at Mactan-Cebu International Airport.

(A) Off-peak periods

(B) Peak Period

We also calculated the economic benefits resulting from the increase in foreign tourists following the terminal’s opening and demonstrated that these benefits are sufficient to recoup the airport construction costs within 10 years. This method is extremely useful for conducting detailed evaluations of the effectiveness of tourism infrastructure and enables highly accurate analysis even in regions where tourist numbers fluctuate significantly by season.

As such, daytime satellite imagery provides a wealth of useful information regarding tourism. I believe that, with the right approach, it is possible to obtain a wide variety of information beyond just vehicle counts, as in our study. For example, the value of urban green spaces could be an interesting application. Research already exists that uses satellite imagery to identify urban green spaces and analyze their impact on a city’s value. Such information will be crucial when considering a region’s appeal.

With the Basic Plan for Promoting Japan as a Tourism Nation approved by the Cabinet and the Japanese tourism industry booming as the country positions itself as a tourism destination, many accompanying issues are also emerging. While government statistics are certainly useful for analyzing these issues, it is also important to grasp the situation with greater immediacy. In addressing such issues, data held by private companies—such as smartphone GPS data, credit card transaction information, and POS data, as introduced at the beginning—as well as satellite imagery data like that discussed in this article, are considered highly useful.It is important to utilize this new data to formulate appropriate, data-driven policies and measures.