Point color and symbol size based on different variables in mapview

Question

I'm trying for a divergent scale theme in mapview to help visualize gains vs. losses, with:

• point symbol circle sizes on an absolute value scale (to highlight losses as much as gains)
• a divergent color scale fill for the circles (say dark blue>blue>white>red>dark red for most negative>negative>zero>positive>largest)
• mouse over hover label of the original value retained

any ideas?

---

library(tidyverse)
library(mapview)
library(sf)

lat <- rep(34,16)
lon <- seq(-128, -126, length = 16)
value <- c(-1000, -800, -600, -400, -200, -100, -50, 
            -25, 25, 50, 100, 200, 400, 600, 800, 1000)

#make data.frame
df <- data.frame(lat, lon, value) 

#make spatial object for mapview
df <- st_as_sf(df, coords = c("lon", "lat"), crs = 4326) %>%
      mutate(value_abs = abs(value)) #value_abs intended for `cex` argument

pal <-  mapviewPalette("mapviewSpectralColors") #from mapview doc. example
m   <-  mapview(df["value"], #sets hover over value as this column
         cex = "value",      #sets circle diameter scaling on this column
         legend = TRUE,
         col.regions = pal(100), #closest I found to a red-blue divergent scale
         layer.name = "value")  
m

In other words, I'm hoping for the pattern of points here below to be symmetrical with the left side as a mirror image of the right in size, but with blue circles at left, red at right, and still allowing the user to see the actual (non absolute) values (e.g. -1000) by mouseover.

attempts: switching cex = "value" with cex = "value_abs" yields warning: In min(x) : no non-missing arguments to min; returning Inf without any points drawn, or with cex = df$value_abs (no quotes), which makes uncolored, enormous points. I'm not planning on needing two legends - just one for either the circle size or fill, showing a min and max value like it does now, would be great.

Hanahanae

Editing your question to improve it is great, but please try to "merge" your edits into more significant chunks. Making a large number of trivial edits increase the noise volume of the site and can be annoying to our users. Thank you for keeping this in mind going forward! – Sapheaded

Answer 1

You are very close. You need to explicitly refer to df$value_abs. Look below:

library(tidyverse)
library(mapview)
library(sf)

df <- data.frame(lat=rep(34,16), 
                 lon=seq(-128, -126, length = 16), 
                 value=c(-1000, -800, -600, -400, -200, -100, -50, 
                         -25, 25, 50, 100, 200, 400, 600, 800, 1000)) 

df <- st_as_sf(df, coords = c("lon", "lat"), crs = 4326) %>%
               mutate(value_abs = abs(value))

pal <-  mapviewPalette("mapviewSpectralColors")

mapview(df["value"], 
                cex = df$value_abs/100, 
                legend = TRUE,
                col.regions = pal(100), 
                layer.name = "value")  

^{Created on 2019-06-24 by the reprex package (v0.3.0)}

Humblebee

But there no info about the symbol size in the legend? – Elysium

Using cex is somewhat of a hack, isn't it? even using mutate I find it hard to have some balanced-sized points. The difference between the large and small points is too big. – Elysium

@HermanToothrot I don't understand why you calling it a "hack". OP didn't ask for size to be a part of the legend. Please post a new question if you have follow up questions. You can share a link to this post and it's ok to tag me and share a link to that question of yours with me. Cheers. – Humblebee