search

ATFutures / who

Active Transport Futures planning tools for the World Health Organisation
8 stars 2 forks source link

Distribute densities onto OSM nodes #2

Closed mpadge closed 6 years ago

mpadge commented 6 years ago

Duplicate of this popdens issue

Robinlovelace commented 6 years ago

This is now closed as there is population data in "../who-data/accra/osm/nodes_new.Rds"

Robinlovelace commented 6 years ago

Note: this only include population cells that have nodes in them. around 1/3rd in Accra do not - see here:

Robinlovelace commented 6 years ago

image

Robinlovelace commented 6 years ago

@mpadge there are at least 3 options here:

1) Do nothing - say we're only dealing with populations that can be represented as being near the road network (the road net should be more detailed) 2) Fix a - re-assign raster cells without road nets in them to surrounding cells - fast but cannot think of the raster incantation to do that atm 3) Fix b - re-assign raster cell pops to the nearest nodes

I suggest 1, 3, and 2 in descending order of good-ness. Thoughts?

mpadge commented 6 years ago
  1. Def-o, because only that way will we be able to translate flows into absolute numbers
mpadge commented 6 years ago

@Robinlovelace can you pls wrap the code for this in a function in popdens? We want this to be truly one click reproducible

Robinlovelace commented 6 years ago

The data in "../who-data/accra/osm/nodes_new.Rds" should be sufficient to test the population-weighted routing. That would be sufficient for the minimum viable product. Working on the function now but not sure it's the best use of time in the context to tight deadlines.

Robinlovelace commented 6 years ago

Closing for now based on the agreement that 'fast and dirty' is OK for now but that fast and clean is better in the future.