Data contains the pixel-level attribute statistics required for regression modeling. Experimental area (Wuhan) is divided into pixels with 500m (308 x 250), 200m (770 x 625) and 100m (1540 x 1250) resolution respectively. In each gridxxx.csv, the fields used are explained in the following table.
| Field | Description | Field | Description |
|---|---|---|---|
| count_poi1 | Leisure and entertainment | count_poi2 | Accommodation |
| count_poi3 | Parking lot | count_poi7 | Medical service |
| count_poi8 | Hospital | count_poi11 | Residential community |
| count_poi14 | Government agency | count_poi18 | Auto service |
| count_poi21 | Research and education | count_poi22 | Shopping |
| count_poi23 | Financial services | count_poi24 | Restaurant |
| building_area | area of building patch data | mobile_night | counts of mobile positioning data |
| sub_id | id of the street that pixel belongs to | county_id | id of the district that pixel belongs to |
- Python 3.x
- You need to import numpy, pandas and sklearn
In populationSpatialization.py, you need to set RESOLUTION, N_ROW and N_COLUMN first for choosing pixel resolution, and then you can run populationSpatialization.py to implent population estimation.
Result contains the predicted population of three resolutions (500m, 200m and 100m).The id and the predicted population of the pixel are listed in the file.
We have reproduced Ye's model and generated population spatialization results based on the aforementioned modeling data in Wuhan. The code and data are also available in this file.