Plotting points over existing bathymetric/topographic plot?

Question

I'm trying to add a second set of data, as geographic points, to a ggmap plot. I already have bathymetric/topographic data applied to the map using geom_contour and geom_raster functions. But every time I add the function geom_point, I get this message: "Discrete value supplied to continuous scale"

This is what my geographic coordinates data frame (Coords) looks like:

Long      Lat                             Type
1  155.5910 19.93401               Not geo-referenced
2  155.4998 19.83302               Not geo-referenced
3  155.2000 19.52440               Not geo-referenced
4  155.6302 20.00934               Geo-referenced
5  155.6623 19.81197               Not geo-referenced
6  155.5619 19.88102               Geo-referenced

How I obtained the bathymetric data (NOT relevant to this question)

bathydata <- getNOAA.bathy(lon1 = -180, lon2 = -154.5, lat1 = 29.3, lat2 =7, resolution = 7, antimeridian=FALSE)
bathydata <- as.matrix(bathydata)
class(bathydata) <- "matrix"
result1<- as.data.frame(bathydata)
result2<- rownames_to_column(result1, var = "lon")
result3<- gather(result2, lat, value, -1)
finalBathy<- mutate_all(result3, list(as.numeric))

This is what my bathymetric data frame (finalBathy) looks like:

lon       lat value
1     -179.9417  7.058333 -5697
2     -179.8252  7.058333 -5582
3     -179.7088  7.058333 -5322
4     -179.5924  7.058333 -5708
5     -179.4759  7.058333 -5791
6     -179.3595  7.058333 -5781

My code

baseArchipelago = get_map(location = c(-180, 7, -154.5, 29.3), zoom = 6, maptype = "terrain")

mapArchipelago <- ggmap(baseArchipelago) 

MapRaster <- mapArchipelago 
            + geom_raster(data = finalBathy, aes(x = lon, y = lat, fill = value)) 
            + geom_contour(data = finalBathy, aes(x = lon, y = lat, z = value),
                bins = 8, colour = "darkslategray", lwd=0.25) 
            + geom_point(data=Coords, aes(x = -Long, y = Lat, fill = Type, shape = Type), color = "black",
                cex = 1.2, show.legend = FALSE) 
            + scale_shape_manual(values = c(23, 23), labels = c("Point 1", "Point2"), name = NULL) 


MapFinal <-   MapRaster + coord_cartesian()

((Note: without coord_cartesian(), I get the error message: "Error: geom_raster only works with Cartesian coordinates"))

If I remove fill=Type from the aes argument in geom_point, the problem is solved and the points get mapped correctly: But I need the points to be filled by their Type!

And if I add scale_fill_manual(values=c("red", "blue"), labels=c("Point 1", "Point 2"), name=NULL) to the code, the points go away and I get the same message again: "Discrete value supplied to continuous scale"

Looks like your fill aesthetic is already mapped to a continuous value in geom_raster. Why not map your aesthetic using color instead of fill in geom_point? — Z.Lin
So once fill is mapped to one function, it can't be applied to another? I thought, since geom_raster and geom_point use different data sets, then they would be applying the fill to different points? (Changing from fill to color did not work) — Joseph
You can if the mapped variables are of the same type. But from what you've posted, it seems like one is continuous and the other is categorical. — Z.Lin
Is there a way to separate them so that one fill is mapped to the bathymetric data and a different fill is mapped to the geographic points? — Joseph

Ekatef Ekatef · Accepted Answer · 2019-06-10T10:30:46

You may add additional discret scales to represent the discrete variables:

MapRaster <- mapArchipelago +
            geom_raster(data = finalBathy, aes(x = lon, y = lat, fill = value)) +
            geom_contour(data = finalBathy, aes(x = lon, y = lat, z = value), 
                bins = 8, colour = "darkslategray", lwd=0.25) +         
            geom_point(data=Coords, aes(x = -Long, y = Lat, color = Type, shape = ype), 
                cex = 1.2, show.legend = FALSE) +
            scale_discrete_manual(aesthetics = c("colour"),
                values = c("red", "yellow"), 
                labels = c("Point 1","Point2"), 
                name = NULL) +
            scale_discrete_manual(aesthetics = c("shape"),
                values = 16:17, 
                labels = c("Point 1","Point2"), 
                name = NULL)

Which gives

Does it answer your question?