Challenge #113: Interpolating Spatial Grid Values

[ChristineB]

So many different solutions to last week's Challenge have been posted here!

This week's Challenge riffs off a process typically used in spatial analysis with raster inputs: interpolation.  In this Challenge, you are provided with two inputs: a polygon representing the island of Maui (hey, there's still snow in the forecast here in CO...can you blame me for picking a beachy location?) and a table of values representing the elevation measurements for 500 m x 500 m grid cells (much like a Digital Elevation Model).  However, some grid cells contain a value of 0.  We'll use some spatial tools to interpolate, or estimate, the values of the cells containing 0 from a "nearest neighborhood" or surrounding cell values. 

First, build a 500 m x 500 m grid around the island of Maui.   Then, interpolate the missing value using the average of the known measurements from the surrounding cells, or "neighborhood".  Use a neighborhood of the 8 nearest surrounding cells in a unique cardinal direction (see example below; a neighborhood of a cell containing a 0 is outlined in blue. In this example, the new interpolated value of the center cell would be 61.5). Should a missing value be located on the edge, use only the nearest cells in a unique cardinal direction, even if 8 values are not used for the calculation.

Hint: Grid cell Grd37_68 is column 37, row 68. The Grid tool starts with column 0.

[danilang]

A non-spatial solution for me.  I'll have to come back and try the spatial route

Spoiler

Dan

[KOBoyle]

Solution attached.

Spoiler

[olivia_paquot]

Fun one :-)

I use a buffer around the 0 values & then spatial match (touch or intersect), to avoid problem when the missing value is located on the edge (even if not the case in this exercise)

[pasccout]

Here is my solution

[jamielaird]

Here's my solution - enjoyed this one.

Spoiler

Not sure if I made things harder for myself by not using the spatial tools, but I didn't see the value in this case.



Approach based on the below rough notes made when I first examined the problem:



Always feels like an easier task once I've got the logic written out, but maybe that's just the way my brain works!

[JoshuaGostick]

That was a very interesting and unique challenge!

Spoiler

[kelly_gilbert]

This was a fun one, and not too tough. It had a little bit of spatial, a little bit of data prep, and a little bit of data reshaping!

Spoiler

I was going to do this using sequential row and column numbers, but figured that I should use a spatial tool since this is a spatial challenge!

[RichoBsJ]

Hi! Here my solution

[ipeng]

My solution