[R-sig-Geo] Clustering spatial points in a SpatialDataFrame object

[ASANTOS]

Dear r-sig-geo Members,

I'd like to calculate the centroid of each cluster (considering points 
with < 10m distance to the same cluster) of spatial points using mean 
operation (spdplyr package) for coordinates and another operation (sum) 
for the attribute(area) without success.

In my example:

#Packages
library(sp)
library(maptools)
library(spdplyr)
library(cluster)

# Small sample (40 points)
small.sample<-read.csv("https://raw.githubusercontent.com/Leprechault/trash/main/sample_points.csv")

#Convert to spatial object
xy.small.sample <- small.sample[,c(1,2)]
spdf.small.sample <- SpatialPointsDataFrame(coords = xy.small.sample, 
data = small.sample,
proj4string = CRS("+proj=longlat +datum=WGS84 +ellps=WGS84 +towgs84=0,0,0"))

# Transform to UTM
utm.wgs.84 <- "+proj=utm +zone=21 +south +datum=WGS84 +units=m +no_defs"
small.sample.utm <- spTransform(spdf.small.sample, utm.wgs.84)
#

#Convert each cluster to one point
Rc=10 #Maximum distance between points is 10 meters
small.sample.utm.dm<-spDists(small.sample.utm) # Distance matrix
clusters <- as.hclust(agnes(small.sample.utm.dm, diss = T)) 
#Hierarchical Clustering
small.sample.utm using data$class<- cutree(clusters, h=Rc) # Cut into groups 
each 10 meters

# Average of x and y coordinates and area using spdplyr package
small.sample.utm.classification<- small.sample.utm %>%
group_by (class) %>%
summarise 
(area_clu=mean(area),x_clu=mean(coords[,1]),y_clu=mean(coords[,2]))

#Error: Problem with `summarise()` input `x_clu`.
#i Input `x_clu` is `mean(coords[, 1])`.
#i The error occurred in group 1: class = 1.


Here I believe that this error is a cause of the spatial data frame 
attributes have individual names and coordinates not. If I try to use 
something like coords[,1], it doesn't work!!


My goal is:

# Original points representation
plot(small.sample.utm, pch=16)

# Center of the centroids representation
points(small.sample.utm.classification$x_clu,small.sample.utm.classification$y_clu, 
col="red")

# Labelling the area
text(small.sample.utm.classification$x_clu ~ 
small.sample.utm.classification$x_clu, 
labels=small.sample.utm.classification$area_clu, cex=0.9, font=2)

Thanks in advanced,

Alexandre