Open wmjolly opened 2 years ago
The desired result would be a consolidated JSON file with the simplified results.
The example analysis JSON contains multiple features.
Should I plan on generating stats for each one of them?
Or would every feature/geometry be a separate zonal stats request?
I would like just one final stats file that contains stats for each feature/geometry.
On Fri, Feb 4, 2022 at 12:04 PM alexander-petkov @.***> wrote:
The example analysis JSON contains multiple features.
Should I plan on generating stats for each one of them?
Or would every feature/geometry be a separate zonal stats request?
— Reply to this email directly, view it on GitHub https://github.com/alexander-petkov/wfas/issues/53#issuecomment-1030269139, or unsubscribe https://github.com/notifications/unsubscribe-auth/AA4G3D2CLQZ453YLGLSZ3ADUZQPNFANCNFSM5NKNJ36Q . You are receiving this because you authored the thread.Message ID: @.***>
Couple of points to clarify:
On Tue, Mar 15, 2022, 02:30 alexander-petkov @.***> wrote:
Couple of points to clarify:
- WildfireSafe shows min, median, and max stats for each zone, while this ticket specifies min/mean/max. Median is needed, and not Mean, correct?
Yes, Median is needed but just to be safe let's include the mean as well, so: min, median, mean, max for all variables
- WildfireSafe lists histograms for Elevation and Slope, in addition to histogram from Landform. Are Elevation and Slope Histograms also needed?
Yes, histograms for elevation and alope are needed
- WildfireSafe shows each histogram with 5 buckets, instead of counts for each distinct value. How should the WPS represent the histogram?
It would be best if we binned the histograms like WildfireSAFE. So, 5 bins evenly spaced from the min to the max
— Reply to this email directly, view it on GitHub https://github.com/alexander-petkov/wfas/issues/53#issuecomment-1067700176, or unsubscribe https://github.com/notifications/unsubscribe-auth/AA4G3DYPEHDAZ2EEWOFCSZ3VABDDBANCNFSM5NKNJ36Q . You are receiving this because you authored the thread.Message ID: @.***>
Added ERC, BI, and SFDI to the zonal_stats CGI script as requested. These rasters are at much coarser resolution (~2.5km) compared to Elevation, Landform, and Slope rasters (30m). Therefore, statistics and histogram strategy is changed to include pixels which are partially within a vector zone, even if the center of the pixel does not fall within.
The same strategy applies for all rasters.
Relevant rasterstats documentation: https://pythonhosted.org/rasterstats/manual.html#rasterization-strategy
Make the zonal_stats service "region-aware", and not just CONUS-centric.
I'd like to streamline the way we sample lots of rasters all at once for an analysis region.
Let's start with the Terrain metrics from WildfireSAFE:
Elevation and Slope should come from the LANDFIRE layers.
Example analysis vector shared as GeoJSON by email.