How can I match coordinates with their associated values in R-CodePudding

I have 250 points that I generated within a rectangle (-4,4)x(-6,6). If the popints are within a certain space they are blue and if they are outside of that space they are red.

The code I used for this is here, where i defined the confined space with squares:

library(sf)

border <- matrix(c(
-6, -4,
-6, 4,
 6, 4,
 6, -4,
-6, -4
 ), ncol  = 2, byrow = TRUE) |> 
 sfheaders::sfc_polygon() 

# sample random points
rand_points <- st_sample(border, size = 250)  

squares1 <- matrix(c(
 -4, 0,
 -4, 3,
 -1, 3,
 -1, 0,
 -4, -0
 ), ncol  = 2, byrow = TRUE) |> 
  sfheaders::sfc_polygon() 

 squares2 <- matrix(c(
  -2, -4,
  -2, -1,
  1, -1,
  1, -4,
  -2, -4
 ), ncol  = 2, byrow = TRUE) |> 
 sfheaders::sfc_polygon()

squares3 <- matrix(c(
 2, -2,
 2, 1,
 5, 1,
 5, -2,
 2, -2
 ), ncol  = 2, byrow = TRUE) |> 
  sfheaders::sfc_polygon()

 squares <- c(squares1, squares2, squares3)
       red_vals <- st_difference(rand_points, squares)
       blue_vals <- st_intersection(rand_points, squares)`
 plot(border)
 plot(negative_vals, add = TRUE, col = "red")
 plot(positive_vals, add = TRUE, col = "blue")

My goal is to match the points' coordinates with their expected value. Example:

In the table, the third coloumn is for the blue points and the fourth coloumn for the red. If the point at that coordinate is blue it gets a 1 and if it is not blue at that coordinate -1, and vice versa for the red points.

So far, I have attained the coordinates of all the points.

        y <- c(red_vals)
        x <- c(blue_vals)
        cdata <- c(x, y)
        coord <- st_coordinates(cdata)`

I am now stuck on trying to figure out how I can classify x and y to their respective coordinates and indicate this in a dataframe.

Any help is appreciated.

CodePudding user response：

You could do:

red_vals <- rand_points[rowSums(st_intersects(rand_points, squares, F)) == 0]
blue_vals <- st_intersection(rand_points, squares)

df <- rbind(cbind(st_coordinates(red_vals), PosGroup = 1, NegGroup = -1),
      cbind(st_coordinates(red_vals), PosGroup = -1, NegGroup = 1)) |>
      as.data.frame()

head(df)
#>             X           Y PosGroup NegGroup
#> X1 -5.2248158  0.03710509        1       -1
#> X2 -5.8932331 -1.41421992        1       -1
#> X3 -0.0609895  0.26541100        1       -1
#> X4  1.7345333 -3.04312404        1       -1
#> X5 -4.6801643  0.24656851        1       -1
#> X6  1.3190239  3.36491623        1       -1

Obviously the first few values are all red dots.

We can see that the points are correct by using this data frame to draw points in ggplot:

library(ggplot2)

df %>%
  ggplot()  
  geom_sf(data = squares)  
  geom_point(aes(X, Y, color = factor(PosGroup)), pch = 1, size = 3)  
  theme_classic()  
  scale_color_brewer(palette = "Set1", direction = -1)