# Merging 2 non overlapping points datasets

I have 2 dataset of points, each one measures a different variable, i need to "merge"(not sure if it is the correct term) these two into one dataset.

My first instinct is to take a purple point find k closets green points and assign the average as the green dataset value for the purple point.

Example image of the points.

Is there a recommend method for this kind of problem?

Edit:

I am using R, but if there is not package that solves the problem i will be fine with a theoretical reference.

Edit 2:

Added code to generate an example, i would like a to know which methods are available and which ones are recommended for a case like this.

``````library(sf)

N_points <-  100
X_range <- c(0, 20)
Y_range <- c(0, 20)

set.seed(123)
df_1 <- data.frame(
x = truncnorm::rtruncnorm(n = N_points,  a=X_range[1], b=X_range[2], mean=10, sd=5),
y = truncnorm::rtruncnorm(n = N_points,  a=Y_range[1], b=Y_range[2], mean=10, sd=5),
value = rnorm(n = N_points, mean=7, sd=2)
)  %>%  st_as_sf(coords=c("x", "y"))

set.seed(753)
df_2 <- data.frame(
x = truncnorm::rtruncnorm(n = N_points,  a=X_range[1], b=X_range[2], mean=10, sd=5),
y = truncnorm::rtruncnorm(n = N_points,  a=Y_range[1], b=Y_range[2], mean=10, sd=5),
value = rnorm(n = N_points, mean=10, sd=2)
)  %>%  st_as_sf(coords=c("x", "y"))

plot(st_geometry(df_1), col="purple", axes = TRUE)
``````

Edit 3, my description of the problem was not clear enough:

Example:

Each purple point represent a measure of air pollution.

Each green point represent a measure of radiation.

I want as result a datasets that has both a measure of radiation and a measure of air pollution for each point.

• What GIS software are you using? If your question is theoretical then please make that clear and tag it with the gis-principle tag.
– PolyGeo
Commented May 1, 2021 at 4:20
• Are you looking for an implementation of that "k nearest neighbours + averaging" thing specifically, or do you want a more general discussion of how to relate two point data sets? It would help if you could supply some data or show some code to generate a simple test data set so we can all work on it. Commented May 1, 2021 at 6:21
• I'm not sure what class of general problem you are trying to solve. Taking the k-nearest and averaging is one thing, it does what it does. Quite what it means in terms of your underlying problem I don't know. What's the big question? How do these two sets of points relate in the real world? Commented May 1, 2021 at 13:00
• @Spacedman, i never had this problem before so i am not sure if that is a good idea or there is "recommended way" to do it. Commented May 2, 2021 at 5:09
• What I don't understand is what your problem is. You have two measurements of different variables and want to combine them into a single set of measurements of something? What's that "something" going to be? Suppose the first measurement is percentage of gold in a soil sample, and the second is percentage of silver - how does it make sense to merge them? Commented May 2, 2021 at 9:12