Hi,
I first posted a question yesterday which asked how to set a formula to calculate distances (by using a function I provided) among all the observations (each one with a latitude and longitude value expressed in decimals) in an excel sheet. Here it is the previous thread.
I slightly modified the solution MarvinP suggested to me by using my own function to calculate distances among coordinates.
Attached you can find the excel with this. You will see that on the right side of the sheet, in yellow background, there is a matrix whose non-blank cells each one reporting a case of close coordinates (e.g.: 4-5 means that observations 4 and 5 are less than 15 Km apart one from another). Cells of the left-side matrix contains the function used to calculate distances among each observation with any other. Matrices are 116X116.
However, I am now jammed because I don't find an easy way to pool the observations found to be close each other (I have used a distance threshold of 15 Km - you can see it in the formula inside each cell of the "yellow" matrix").
The ideal way to establish new coordinates values for observations close one another would be by using coordinates which are at an intermediate position, however, this is not my priority, it would be OK also using one arbitrary value for all of them, i.e. picking up one pair of coordinates of one of the close observations and attribute this to the other. I have seen that, by excluding nonsensical similarities (e.g. 1-1, 2-2, etc.), there are groups of close observations of size 2 (e.g. 4-5) up to 10 (nine observations very close to a given one).
I have thought it for a while but it doesn't seem I am too inspired.
Anyone of you willing to solve this puzzle?
Simone
Bookmarks