less than 1 minute read

Traveltime Maps, Population Density estimation.

All areas that are reachable within a certain time from a location are called isochrones or Route Range.

  • There are a lot of public APIs available to visualize these regions but most of them want some $$$. I have made a short compilation of them in the doc for the datasciencedojo intern thing.

Population Density estimation is also a real problem since census aren’t usually reliable or very high resolution. There are a few organizations that have used satellite imaging and other geographical data with some machine learning algorithms to get a better estimate of population density than a census and also with a higher resolution (~30m). The 2 main guys I read are Facebook(High Resolution Settlement Layer: HRSL) and WorldPop.

  • Facebook used Deep Learning with ResNet and other Data
  • WorldPop used Random Forest Regression tree based mapping and other Data.
  • There is this thing “Dasymetric” which these people mention in their methods that I didnt bother looking into.
  • OpenStreetMap is a thing..
  • GIS is geographic information system/science