Calculating area of pixles for certain Aspect value per specific area?

[SawsanHadi]

I have the aspect (raster file) for the roof of buildings at the city scale. Now I want to calculate the area of each aspect group per building. I tied zonal statistics, zonal geometry and etc . Also reclassified the aspect layer and tried to add the Building ID. But I didn't get the desired result. Is there a way to do that? Kindly see the attached

[DanPatterson_Retired]

to do it on a per building basis you need the individual 'zones' which can be accomplished using.

Region Group—Help | ArcGIS Desktop 

the resultant would be your zone file, then a zonal count would give you the number of cells (times cell area) per aspect by building

[DanPatterson_Retired]

You should also read this blog post...

https://community.esri.com/community/gis/analysis/spatial-analyst/blog/2018/07/05/getting-the-most-o... 

[SawsanHadi]

Thanks Dan! I performed the region group, then displayed with the link value (shows exact values as for aspect.tif). Also did zonal statistic with the feature zone as the regiongroup.shp + the zone filed as count+ value raster (uploaded aspect.tif).  But still didn't get the desired result . I think I am missing something , can you please illustrate the second step in more detail.  

[DanPatterson_Retired]

need to see what your desired result is in light of what you have.. it may need  an interpretation suggestion or modification

[SawsanHadi]

After visualizing the product of Region Group , it seems that it is not giving the zone needed (top classification by value bottom classification by Link).

 

[DanPatterson_Retired]

something doesn't look right.  You need to regiongroup your original zones.  Then you use that for the zonal statistics on your aspect.  You don't regiongroup the aspect

[SawsanHadi]

Thank you for your time! I do agree with you, but I did regiongroup for the building footprint first but it didn't give what I needed. That is why in the second attempt I did the regiongroup for Aspect. It is worth to mention that both layers follow the same grouping pattern when I reclassify the value, with the exception of the No. of elements in the attribute table -which are different- .

[DanPatterson_Retired]

regiongroup produces the zones for zonal statistics operations.

now you need to do the zonal statistics

I notice that you have an incredibly large number of zones, suggesting that your input raster wasn't cleaned up and you probably have a whole load of zones that consist of 1 or 2 cells.  You could use nibble to fill those in with adjoining values, but I suggest that it might be worthwhile getting a clean vector file of the building footprints and start the zonal statistics from that