How to bin x,y,z vectors into matrix (R)

问题

This question was migrated from Cross Validated because it can be answered on Stack Overflow. Migrated 5 years ago.

This is probably a basic question, but I haven't been able to Google anything helpful after trying for days.

I have an R dataframe with x,y,z tuples, where z is a response to x and y and can be modeled as a surface.

> head(temp)
         x         y        z
1 36.55411  965.7779 1644.779
2 42.36912  978.9721 1643.957
3 58.34699 1183.7426 1846.123
4 53.55439 1232.2696 1990.707
5 50.76167 1115.2049 1281.988
6 51.37299 1059.9088 1597.028

I would like to create a matrix of mean z values, with rows representing binned y values and columns representing binned x values, like

            0<x<40    40<x<60  60<x<80   x>80
    0<y<800 1000.0    1100.00  1100.00   1000.0 
 800<y<1200 1000.0    1200.00  1200.00   1000.0 
1200<y<1400 1000.0    1200.00  1200.00   1000.0
     y<1400 1000.0    1100.00  1100.00   1000.0 

thanks

回答1:

You could use tapply and cut here

with(temp, tapply(z, 
    list(
      y=cut(y, breaks=c(0,800,1200,1400,Inf), include.lowest=T), 
      x=cut(x, breaks=c(0,40,60,80,Inf), include.lowest=T)
    ),
    mean)
)

The cut function splits x and y into what ever groups you like. While tapply calculates the mean here for each grouping of the those values in the list.

#                    x
# y                     (0,40]  (40,60] (60,80] (80,Inf]
#   (0,800]                 NA       NA      NA       NA
#   (800,1.2e+03]     1644.779 1592.274      NA       NA
#   (1.2e+03,1.4e+03]       NA 1990.707      NA       NA
#   (1.4e+03,Inf]           NA       NA      NA       NA

来源：https://stackoverflow.com/questions/24841888/how-to-bin-x-y-z-vectors-into-matrix-r