Blending the surface and scatterplots in one 3D plot

Question

I study patterns of whale distribution around specific seabed structures. I am trying to create an interactive 3D graph at the same time:

• bathymetry as a surface ( x = longitude, y = latitude, z = depth) and
• geographic location of whale groups ( x = longitude, y = latitude, z = fixed depth of -30 meters, for example).

Coordinates are projected into the UTM coordinate system.

I usually work with R and the ggplot2 package to create shapes. Here the plotly package looked like a good option.

I started with the bathy_ras bathymetric raster and the data.frame points .

 > bathy_ras class : RasterLayer dimensions : 784, 821, 643664 (nrow, ncol, ncell) resolution : 102, 111 (x, y) extent : 755070, 838812, -2612148, -2525124 (xmin, xmax, ymin, ymax) coord. ref. : +proj=utm +zone=58S +datum=WGS84 +ellps=WGS84 +towgs84=0,0,0 data source : in memory names : bathymetry values : -1949.42, -34.27859 (min, max) > str(points) 'data.frame': 214 obs. of 3 variables: $ x: num 774264 777293 775476 773430 773284 ... $ y: num -2534165 -2533556 -2531012 -2532904 -2533695 ... $ z: num -30 -30 -30 -30 -30 -30 -30 -30 -30 -30 ... 

I cannot find a way to combine two datasets on the same graph / same axis. I tried two methods, but none gave me the result I wanted.

1) Creating a graph in Rstudio using a graphics package.

 #convert raster into a matrix of bathymetry values bathy_matrix <- as.matrix(bathy_ras) > str(bathy_matrix) num [1:784, 1:821] -362 -365 -367 -369 -371 ... #create interactive plot p <- plot_ly(z=bathy_matrix, type="surface",showscale=FALSE) #this works fine, it get a 3D interactive surface plot of the seabed #but if I try to add the "points" layer it doesn't show on the plot p <- plot_ly(z=bathy_matrix, type="surface",showscale=FALSE) %>% add_trace(x = gp_seamounts_df$utmx, y = gp_seamounts_df$utmy, z = gp_seamounts_df$z, type = "scatter3d", mode = "markers", showscale=FALSE, opacity=0.98) 

---

2) Creating a plot entirely on the site. First, I converted the raster "bati_ras" to a matrix containing all coordinate points (x, y) and depth z

 #convert raster into a dataframe bathy_df <- as.data.frame(coordinates(bathy_ras)) bathy_df$z <- values(bathy_ras) > str(bathy_df) 'data.frame': 643664 obs. of 3 variables: $ x: num 755121 755223 755325 755427 755529 ... $ y: num -2525179 -2525179 -2525179 -2525179 -2525179 ... $ z: num -362 -361 -360 -359 -358 ... 

I created a story account. I imported two data files as .txt files in my graphical account: bathy_df and points .

This creates two grids in the graph. I can easily write two separate 3D plots for these two data.frames: one is a surface plot (shown below), the other is a scatter plot. I tried to include the scatterplot in the surface plot as a new trace following this tutorial ( http://help.plot.ly/update-a-graphs-data/ ), but the insert into option is not available if the scatterplot is in 3D surface graphic created from web GUI

Is it possible to combine a scatter3D and a piece of surface in plotly ?

nb: I tried raster::persp in combination with points() , but I am not very satisfied with the general aesthetics of the surface, so I would prefer to do this with plotly and / or ggplot2 .

Answer 1

I believe that what you are doing should work fine. I think this may be relevant to your x and y coordinates. The surface uses 1:ncol(bathy_matrix) as the x axis and 1:row(bathy_matrix) as the points of the y axis (check marks, if you like).

Your points will need to have x and y coordinates in this range so that they appear on the surface graph. The following is a simple example.

 set.seed(123) x = sample(1:ncol(volcano), size = 50) y = sample(1:nrow(volcano), size = 50) z = c() for(i in 1:50) {z <- c(z, volcano[y[i], x[i]])} df <- data.frame(x, y, z) plot_ly(z = volcano, type = "surface") %>% add_trace(data = df, x = x, y = y, z = z, mode = "markers", type = "scatter3d", marker = list(size = 5, color = "red", symbol = 104)) 

I get this:

Hope this helps ...