Study On Urban Population Data Spatialization In City

Population data is one of the basic data of geography, demography, sociology.And the traditional Census data is collected on the administrative unit, which has low spatial resolution and hard to integrate with other data. This impacts the application on the related research.Research on spatialization of Population data can solve this problem and become an attractive field in geography.However, it's more research on large scale,less on small scale especially in the city.As the goal of obtaining the high spatial resolution population in urban,this paper analysis the difference method of Population data spatialization with different data sources in downtown Fuzhou. Firstly, collect data and establish the Database;Then us the method of density of land use, the method of dwelling unit, and the method based on POI data to get the grid population data. The method of density of land use is based on Residential Land classification by high resolution image, and the population of 100m* 100m grid is calculated by the area of different Residential Land classification in the grid.The method of dwelling unit is based on residential building classification by high resolution image,the population of 100m* 100m grid is calculated by the number of different residential building classification in the grid.For the method based on POI data, analyze the regulation of the Pearson correlation coefficient computed by population and POI at different range (50 m step length) respectively, in the grid of 100 m to 1000 m length.Then, spatialize the population data based on the data of POI in 300 m* 300 m grid. Finally, gridding the building level census data to evaluate the accuracy of these methods. Research on population spatial based on POI is less involved in previous studies, can provide a reference for future study on spatialization based on POI data.This study found that land use density method and dwelling unit method is suitable for small area in small scale, the POI data is suitable for the grid which the length is above 300 m, but the latter is more convenient than the former two. The larger the scale,the better the results based on the POI data; Estimate the population by POI is reliable with the grid length more than300 m, the grid 600 m above is better. The number of POI around grid should be considered, in commonly 500 m-650m radius can get a better results.