colorNumeric with a palette function is 'bleeding' colors

[mtennekes]

Consider this 'binary' palette function:

library(leaflet)

# Example data
df <- data.frame(
	lat = c(40.7, 34.0, 41.9),
	lng = c(-74.0, -118.2, 12.5),
	value = c(10, 20, 30)
)

pfun = function(x) {
	ifelse(x < 0.5, "blue", "red")
}

pal = colorNumeric(palette = pfun,
				   domain = df$value)

When feed into addLegend, it seems to 'bleed' colors:

leaflet(df) %>%
  addTiles() %>%
  addCircleMarkers(
    ~lng, ~lat,
    color = ~pal(value),
    radius = 8,
    fillOpacity = 0.8
  ) %>%
  addLegend(
    "bottomright",
    pal = pal,
    values = ~value,
    title = "Value",
    labFormat = labelFormat(suffix = ""),
    opacity = 1
  )

I was expecting a hard distinction between blue and red in the middle of the value range. Is this a bug, or am I missing something? Triggered by r-tmap/tmap#1175 (comment)

[mtennekes]

Looking at the browser inspector I found out that the SVG gradient only consists of 7 stops. That explains the behaviour. Is there a way to set e.g. 100 stops?

The aim is to create a legend like this:

If there are other ways to do this, I'm happy to know.

[trafficonese]

Maybe you want to use colorBin or colorFactor instead?

For example:

pal <- colorBin(
  palette = colorRampPalette(c("red", "orange","yellow","lightgreen","blue"))(20),
  domain = df$value,
  bins = 20
)

[mtennekes]

Thx @trafficonese The point is that a couple of users suggested to place the labels 10, 11, 12, .. 30 at the cut points, rather than printing the interval 10-11, 11,12, etc. Why: because it is often unclear that 11 is not part to the first interval 10 - 11.