Skip to content

Latest commit

 

History

History

tutorials

Folders and files

NameName
Last commit message
Last commit date

parent directory

..
README.md
README.md
 
 
count_estimation_tutorial.ipynb
count_estimation_tutorial.ipynb
 
 
data_prep_tutorial.ipynb
data_prep_tutorial.ipynb
 
 

README.md

Tutorials

Data Preprocessing Tutorial

In the notebook data_prep_tutorial.ipynb, we provide the code for preparing all the necessary data for estimating the total object count in a region.

The notebook will prepare following data are needed to run IS-Count:

  • A binary mask with the same resolution as the covariate raster for the region of interest
  • An "all-pixel" file that contains [all_lats, all_lons, all_s_pix, all_probs['nl'], all_probs['pop']] of the region of interest

Before running the notebook, make sure you have all the required data downloaded and have a directory under the root path named sample_data/.

  • The covariate raster data could be downloaded here.
  • The Microsoft Building Footprints data is available to download here.
  • The country-wise Google Open Buildings data is available to download here.

The sample_data/ folder needs to be structured as follows before you run the code in the notebook:

sample_data
├── covariates
│   ├── NL_raster.tif
│   └── population_raster.tif
├── ms_building_footprint
│   ├── us
│   │   ├── NewYork.geojson
│   │   └── ...
│   └── ...
├── open_buildings
│   ├── us
│   │   └── ...
│   └── ...
└── shapefiles
    └── us_states
        ├── cb_2018_us_state_20m.cpg
        ├── cb_2018_us_state_20m.dbf
        ├── cb_2018_us_state_20m.prj
        ├── cb_2018_us_state_20m.shp
        └── ...

Count Estimation Tutorial

In the notebook count_estimation_tutorial.ipynb, we provide the code for actually estimating object count in a given region, using the data we prepared in the previous step.