you can also produce a kernel density layer and compare the values for the two clusters. The kernel density of course will be calculated relative to the whole set, so values will change as you change the set of points used to calculate your kernel density surface. There is a tool in ArcGIS for it I am not sure in QGIS, but I'd assume there would be an equivalent.

are you sure you are looking at the right output? it looks like you have the standard deviation map rather than the mean projection?

its better if you post this to a more suited SE site like Open data...opendata.stackexchange.com

great! I agree sometimes its a bit hit and miss. if I am not mistaken tiff can handle bigger sized files better (compared to the standard jpeg format)

I don't think the format is an issue, although I always prefer using tiff..

for example if the road id of the one you d not want to label is 5. Then you can use [roadId] <> 5. will label every road except those that have id # 5. For string you can use not([roadname] = 'Main') for example.

Most tiff or img files come with a pre-set value to denote no data either a very large -ve number or -9999 either way that will be noted in either the metadata or the header of the raster.

I don't understand what you mean. if you run it without a setnull and it gave you a negative number but when you run the setnull it gives you a small positive number. It means that the setnull did it's job and set all -ve values to no data. one way to check is to do a - Con(( IsNull(data),-1,data) so that now you can set all your no data back to constant -1. The -1 values will show you all the cells the setnull function set into no data. In ArcGIS It is always recommended to do raster calculations on ESRI GRID anyway, so convert any images to that format before any spatial analysis.

The number you gave above [9.53674e-007] is not a negative value it is a scientific number form of 0.000000953674. ( you can copy it to excel and re-format it to general number). So which means you have a very small positive value that is greater than zero. So setnull is doing its job properly.

Ha ha we posted almost at the same time, I'll put mine as a comment :)

did you check if you have access to editing your data first? I mean is it on a server with no write access to you, or is it in a read-only folder?. If you have access. Next did you use Arc Toolbox > Analysis Tools>Overlay toolbox or another set of tools?