National grid of values for modelled change in cycle counts

[joeytalbot]

Predictions resulting from a GAM model of national DfT counts, for the years 2010 - 2019 on a 500m grid:

https://github.com/saferactive/saferactive/releases/download/0.1.2/gam-full-results-peak-grid-national.Rds

[Robinlovelace]

To convert to LA level and check-in with @mem48 on Weds.

[mem48]

@joeytalbot for GAM model results do not cover the whole of the UK

The edge of Cornwall and Lowestoft have been cut off

[joeytalbot]

I'm sorry about that I may need to rerun the model and expand the grid bounds slightly.

[Robinlovelace]

Just a follow-up on this, I don't think this is a blocker. Are you still working on this @mem48 or is it stuck on #57 ?

[mem48]

@Robinlovelace having internet problems, but I got this working.

[Robinlovelace]

Looking good @mem48. Any ideas of similarities/differences compared with this?

Reproducible example below. We now have 2 raster_rnet_bicycle.tif files it seems although as far as I can tell only 1 of them has been released and is based on reproducible code that has been committed to this repo.

Code that may be useful for comparison:

u = "https://github.com/saferactive/saferactive/releases/download/0.1.1/raster_rnet_bicycle.tif"
# download.file()

lnd = spData::lnd
lnd_20km = stplanr::geo_buffer(lnd, dist = 20000) %>% 
  sf::st_transform(27700)

r = raster::raster("~/saferactive/saferactive/rasters/raster_rnet_bicycle.tif")
r_lnd = raster::crop(r, lnd_20km)
raster::plot(r_lnd)

[mem48]

I don't know about the raster, but what I'm doing is based on the GAM model.

[Robinlovelace]

I don't know about the raster, but what I'm doing is based on the GAM model.

So what is this? #75 (comment)

I thought based on your comments in #57 you were working on re-splitting-up the network as the basis of a raster map of cycling, although I would suggest that the results in the map above are sufficient for our purposes.

Good to focus on the GAM model.

This issue, as I understanding it, is about multiplying the values in the pre-existing raster visualised above by the results from the GAM model. Any updates/clarifications welcome.

[Robinlovelace]

Nudge on this @mem48 - I suggest writing code that works with the raster produced by @joeytalbot - who I think plans to update it (please let us know when you're working on it and keep us posted). Writing the code will probably take longer than running so think it's still a strong basis for moving forward on this. Will be great to see the per-year raster outputs.

[Robinlovelace]

Will be good to get closure on this issue. I think a summary of the latest data on grid values for cycle commuting in 2011, and equivalent data showing multipliers and KSI will be enough to close it. I think the data is already there but we just need to document it. Please take a look when you get a chance @joeytalbot.