What is in between?

Going through this exercise is even more up to you than usual.

Areal interpolation

Pick any dataset linked to polygon geometries used in previous chapters of the course. Or any of your own, if you wish.

• Create a hexagonal grid of a reasonable resolution covering the area of your data.
• Interpolate data to the grid. Make sure to use extensive and intensive variables correctly.
• Do you understand the difference in how the intensive variables are interpolated vs. the extensive?
• If you don’t have a categorical column in your data, create one. Interpolate this categorical column to the grid. What is different when dealing with categorical data?

Point interpolation

When dealing with point patterns, you have worked with Airbnb listings in Prague. Use the same data to interpolate the price of a single-bedroom flat to a hexagonal grid of resolution 10.

Generating H3 polygons

You will need to pass a GeoSeries of polygons to the h3fy function. You can either get those from the Spatial weights session or creating a GeoSeries on-the-fly.

Creating geometry on-the-fly

This would be one option:

extent = gpd.GeoSeries([airbnb.union_all().convex_hull], crs=airbnb.crs)

Data cleaning

Don’t forget to remove outliers. Note that the price is now in CZK, not in GBP, as it was in the case of Edinburgh.

• Play with different interpolation methods and try to understand the differences.
• Test how KNeighborsRegressor results differ when changing the number of neighbours.
• Do you understand the optimal parameters of Kriging?
• Can you figure out how to do Simple Kriging instead of Ordinary Kriging covered in the hands-on section?