Kernel or Point Density output unit points per ... furter explanation needed - radius

Discussion created by peter_m on Nov 27, 2012
Hello froum users!

I want to display the density of house sells in a city, with a ruster surface. Input data are points - one point per sell.

So for this tool I have found tow video tutorials:
(the hole video)
(only the first 16 min.)

So I have started with the fowolling steps:
1) Integrate (to summarize the points in a area)
2) Collect events (to get a count field)
3) Calculate Distance Band (as shown in the video spatial best-practices)
4) Calculate the "search radius" - one of the inputs for the density tool
   - Load "Incremental Spatial Autocorrelation" toolbox (as shown in the video spatial best-practices)
   - Values from Distance Band
      Maximum Distance = Beginning Distance
      Average Distance = Distance Increment
A: Now the peaks show the possible right "rearch radius" for the density input? (right)?

But whats about the optional "output cell size" and the area units ???
- Map unit is meter
- output cell size could be from 50 meter to 100 or 1000
- Area units is also equal for me, but I want to know what I get out. Thats important !!!!

I have looked in the tool help but there is not a good explanation as I think (and my englisch isn´t the best) so I doesn´t understood it!

As you can see in the attachment I get the search radius peaks 1990, 2500 and 3100.

- So if I choose 2500 as "search radius"
B: What "output cell size" should I choose ??
C: What "area unit" schould I choose ??
I know that all three imputs musst be right to get observations per square kilometer or per hectars.

D: Please can anyone explain me waht "output cell size", "search radius" and "area unit" I should choose to get observations per square kilometer or per hectars? (and how this works)

Thanks a lott !!!