binary raster plots with more than 2 colors

[jamiemkass]

When plotting a binary raster with addRasterImage(), even when the parameter "colors" is set to two colors, the plotted raster shows more than two colors. It seems that it's interpolating color values, even though this is not desired with a binary raster. Zoom in on the example below to see. If there's a way to fix this, please let me know.

v <- sample(c(0,1), 40000, replace = TRUE)
x <- raster::raster(nrows=160,ncols=250,xmn=-120,xmx=-80,ymn=15,ymx=40,vals=v)
m <- leaflet() %>% setView(-100, 25, zoom = 3) %>% addProviderTiles('Esri.WorldTopoMap')
m %>% addRasterImage(x, colors = c('red','blue'))

[tim-salabim]

This is because of the internal reprojection of raster layers. See also #219. I don't think this has been addressed yet (@jcheng5 correct me if I'm wrong). For the time being, you can use mapview to set up your map which will reproject using the nearest neighbor method for factorial rasters (you need to convert your layer to factor first).

library(mapview)
library(raster) # needs to be loaded to access as.factor for raster class

v <- sample(c(0,1), 40000, replace = TRUE)
x <- raster::raster(nrows=160,ncols=250,xmn=-120,xmx=-80,ymn=15,ymx=40,vals=v)
x <- as.factor(x) # convert to factor so we know c(0, 1) are not to be treated as numeric/integer
mapview(x, map.types = 'Esri.WorldTopoMap', col.regions = c('red','blue'), legend = TRUE)@map %>% 
  setView(-100, 25, zoom = 3)

[jcheng5]

I would love a PR on this!

[tim-salabim]

I should have time on the weekend

[tim-salabim]

BTW, what's with this one? #229

[jamiemkass]

Thanks for the answer. For my project, it's essential I use the leaflet workflow (although mapview looks very cool), so a change to leaflet would be ideal. If such a change is made, would one simply convert the raster to factorial for an automatic "ngb" resample?

[tim-salabim]

@ndimhypervol once #462 is merged you can do

v <- sample(c(0,1), 40000, replace = TRUE)
x <- raster::raster(nrows=160,ncols=250,xmn=-120,xmx=-80,ymn=15,ymx=40,vals=v)
m <- leaflet() %>% setView(-100, 25, zoom = 3) %>% addProviderTiles('Esri.WorldTopoMap')
m %>% addRasterImage(x, colors = c('red','blue'), method = "ngb")

There is no mechanism to automatically infer which method to use so you need to explicitly set method = "ngb" - the default is method = "bilinear" for backward compatibility.

[jamiemkass]

Great, thanks a lot. Looking forward to the merge!

On Sat, Sep 9, 2017 at 7:29 AM tim-salabim ***@***.***> wrote: @ndimhypervol <https://github.com/ndimhypervol> once #462 <#462> is merged you can do v <- sample(c(0,1), 40000, replace = TRUE)x <- raster::raster(nrows=160,ncols=250,xmn=-120,xmx=-80,ymn=15,ymx=40,vals=v) m <- leaflet() %>% setView(-100, 25, zoom = 3) %>% addProviderTiles('Esri.WorldTopoMap')m %>% addRasterImage(x, colors = c('red','blue'), method = "ngb") There is no mechanism to automatically infer which method to use so you need to explicitly set method = "ngb" - the default is method = "bilinear" for backward compatibility. — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#461 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AB_l29Bb7S0JuJP-r2uH4ZLQYOVSsGBsks5sgna0gaJpZM4PPQ_K> .

-- ----------------------------------------------------------- Jamie Kass PhD Candidate, Department of Biology City College of New York, CUNY Graduate Center Alt email: jamie.m.kass@gmail.com My website: ndimhypervol.github.io Phone: (917) 602 5787 Lab: andersonlab.ccny.cuny.edu