Find Nearest Oddity

[PhilipMannering]

Hey All,

I've found weird behavior in the Find Nearest Tool.

If I look for the two nearest points, I get,

And so, if I look for the 1 nearest I would expect to see Suzuka Circuit. However, I instead see Autodromo Jose Carlos Pace,

Can anybody help explain what's going on? Workflow attached.

[danilang · Accepted solution]

Hi @PhilipMannering 

It appears to be a limitation of the find nearest algorithm.  If you include the direction as an output you see that the direction LV->Suzuka is listed as East where it should be West.  Probably related to crossing the International Date Line.

I added in the site of the Montreal Grand Prix as another site where you'd obviously travel east to get to as well as Saugus Speedway outside of LA as another site to the west.  Finding the nearest 3 then gives

My best guess is that the algorithm uses the direction in some obscure way and hiccups when crossing the IDL.  Maybe a question for the R gurus at Alteryx

Dan

[DataNath · Accepted solution]

Was wondering what was going on with the various Weekly Challenge submissions this week! When I had a peak at a few that got the lower kilometer coverage, none had used Find Nearest. As @danilang suggests, when I plotted mine, it doesn't wrap around the date line as expected but bounces back within the view:

[PhilipMannering]

@danilang and @DataNath . But it does give the correct distance as an output. Just doesn't seem to use that distance when getting the nth nearest points. Guess I should be using Append Fields → Distance Tools in my iterative macro ¯\_(ツ)_/¯