I'm not entirely sure if I understand your question correctly but I think you could use the NeighborFinder for this. Set it to Candidates Only mode.

You could set maximum distance to 100m (or 101, not sure) and generate a list attribute on each point, in which you gather the attributes of interest of the 8 (or less) surrounding points.

You can explode the list attribute using the ListExploder for further analysis/ease of visualisation.

Hope this helps.

It may be easier to do your calculations on the original raster. Have you looked at the Raster transformers?

Thank you for your reply. I have looked on how to do this with raster transformers, like the RasterSlopeCalculator, but it "only" gives the maximum rate of change in z in the cell, and I need to hav the slope for all 8 directions

Thank you fot your reply.

I'm not very familiar with lists. I know the values are like stored and can be called with {} but I don't really know how to retrieve them.

I had already tried the NeighborFinder (Candidates Only mode somhow only returns unmatched elements) but it's not very clear to me what happens

It now seems to work, thank you. I had an error in my formula @sqrt($(Resolution)^2 + $(Resolution)^2) * 1.1, as FME Arimthmetic Editor doesn't understand ^ as power, I have to write @sqrt(@pow($(Resolution),2) + @pow($(Resolution),2)) * 1.1 and everything works fine

Glad to see the NeighborFinder approach works for you. Below is some more explanation about the lists.

This is my dummy workspace:

These are the settings of my NeighborFinder:

As you can see under Attribute Accumulation, I collect all values from the Z attribute to the list called _NeighborList.

You can explode the list with the ListExploder, which results in this:The ListExploder basically multiplies your input features with the amount of list elements. I have prefixed the neighbor attributes with nb_ to avoid confusion. 'Z' and 'ID' are from the original point and 'nb_Z' and 'nb_ID' are from the neighboring points.

As you can see, ID 1 has three neighbors, so three points are generated. These points have the attributes of ID 1 (Z = 3.9) and of the neighbor points ID 2, ID 11, ID 12 (nb_Z = 2.6, nb_Z = 3.2 and nb_Z = 1.3).

You can do further calculations with these exploded features and e.g. merge it back together based on the original ID.

Have fun!

Worked perfectly for me, thanks!