Location issues with Structures

[kkemp85]

I stumbled upon an issue in Arches where some structures are not in their correct locations. It doesn't look like this is an issue with the basemap being offset from the data or anything like that since the structures aren't all in the incorrect locations and they aren't all slightly offset. Some of the structure dot markers seem to be in circular clusters when they should be in a line representing buildings lined up along a street. I first noticed the issue while searching around Flagler College in St. Augustine. Since the entire historic district is in Arches, there are a lot of buildings to view and it seemed like a lot were out of place. For example, the FPAN Northeast office (SJ0083A) is located in tree cover amongst some other Flager buildings, in Arches though, the marker is way South on the street at the end of the block. You can see the actual location compared to the Arches location below.

FMSF View Arches View

This is something I notice throughout the St. Augustine Historic District. Not all locations were off but a lot were. You can see in the Arches image below how they are strangely clustered in certain areas or in the middle of the street. FMSF view Arches view

I tried to do a quick look around Arches to see if I could pinpoint this issue to maybe our most recent FMSF update of sites or if the issue was concentrated in a certain area. From what I can tell the issue isn't as prominent in other areas outside St. Augustine, but I couldn't check every single structure. I noticed a few were incorrect in Hillsborough River State Park, there were a few in Walton county (examples: WA00598, WA00596. WA00571, WA00563, WA00918), Okaloosa (OK01449), Santa Rosa (SR01870, SR01744), and Escambia (ES01833). Some of the ones affected are structures we added in the FMSF update, but I don't know enough about the system to say for sure that all were or that this hasn't been an issue for a while, I just happened to be the first person to notice it.

This could be a big issue considering we have scouts using the system to locate and visit sites and with the markers in the wrong places, it can seem like certain structures aren't in the system even when they are if scouts are using the map to search instead of typing in the name or site ID of a resource to search for it.

[mradamcox]

@kkemp85 All of the geometries in our database come directly from the FMSF, and from the beginning one of our underlying rules was to be a mirror of that database. So we've never changed or updated any of the site records. However, technically there isn't anything stopping you or any admin from going in and changing the geometry for these structures, but personally I think that it would be wise to, at the very least, keep track of these changes somewhere. Down the road it would also be good to add fields to the site resource models to record those changes.

[kkemp85]

@mradamcox I can totally go in to fix these, I guess I'm just wondering what caused the issue in the first place? There are way too many in the wrong locations for this to be something a staff member altered. Maybe something happened in translation to Arches? Because the markers are in the correct locations in the FMSF GIS Data we use to update Arches, but they aren't in Arches. Is there a way to check geometries to get a mass list of which ones aren't in the proper locations?

[mradamcox]

Further investigation: We found that this is due to the fact that the initial resource load from FMSF (in 2017) for these sites did have these inaccurate locations, but since then FMSF has corrected all of the sites. So current FMSF exports show the updated locations. Initial solution for this will be a new management command something like:

python manage.py fmsf update-geometries --input FMSFshapefile.shp

Which would read the siteids and geometries from an input shapefile (which comes regularly from FMSF) and updates the Arches geometries for any resources matching the siteids.

This command should later be expanded to update non-spatial attributes as well, because these get changed now and then in FMSF and HMS should reflect these changes.