Estimation of Distribution of Indoor and Outdoor Population Using Location-Based Serviced Data

It is important to determine the distribution of indoor and outdoor populations when assessing the risk of extreme summer temperatures in urban areas, the heat island effect, air pollution, and disasters. However, conventional indoor/outdoor determination using sensors and survey forms has problems in scale and immediacy, and it is difficult to determine the population distribution by time of day and by subregion. This study developed and verified a new estimation method using location-based service data with the aim of accurately estimating the distribution of indoor and outdoor population in urban areas and clarifying its temporal and spatial variation in some wards of Tokyo Metropolis, Japan.

The research methodology used point-type data from the global positioning system (GPS) communication, grid-type data acquired from mobile phone base stations, and building polygon data, and combined them to estimate the indoor and outdoor population by time and grid. Specifically, the grid-type demographics were used to recover confidential population data from point-type location service data, which were superimposed on building polygons and determined the indoor/outdoor population based on travel distances. The estimated indoor population was compared and validated with the person-trip survey (PT survey) data at the grid level.

As a result, validation with the PT survey in Nerima Ward confirmed that the estimated behavior of the indoor population, including confidential data restoration, was close to the actual behavior. The model is capable of capturing indoor and outdoor population movements by time of day and by grid, and its results are expected to have a wide range of social applications in the future, including risk assessment for heat stroke, air pollution, and disasters.