Use 2014 LA City building data
Closed this issue · 10 comments
Use the new 2014 building data here: http://geohub.lacity.org/datasets/813fcefde1f64b209103107b26a8909f_0
Next steps:
- Merge from attributes from Assessor data.
- Review if data quality issues from 2008 is evident in this version.
- Prepare sample data using the 2014 versions.
Going to take a stab at comparing the footprints tomorrow.
2014 LA City building stats:
- Unchanged - 1,036,741
- Replacement - 6,152
- New - 12,092
- Modified - 54,253
- Total - 1,109,238
Overall quality looks really good. Some observations below:
Building complex tagged as separate adjoining buildings in OSM, map: https://www.openstreetmap.org/#map=18/34.06634/-118.44328
Weird slivers, map: https://www.openstreetmap.org/#map=19/34.24951/-118.28962
Replaced buildings matches 2014-2015 Bing imagery, map: https://www.openstreetmap.org/#map=20/34.06524/-118.44460
New buildings matches 2014-2015 Bing imagery, map: https://www.openstreetmap.org/#map=19/34.14808/-118.28917
Does Pasadena still have the building parts problem #19 ?
@planemad, Pasadena is not part of LA City so I don't think this will be an issue.
On further review, I filtered all data that where modified/replaced/new with an area of <= 100 feet^2 (597 buildings). Out of the 200 buildings I visually inspected, all of them a valid free-standing buildings. I conclude that for the 2014 changes, building as part is not an issue, however, it is possible it is still present in the unchanged data.
I agree. It looks really solid. I'm gonna take a pass at removing the L.A. City buildings from the county file and replacing them with those files.
@jschleuss Thanks for doing the merge! Looking at the sample data now, I'm seeing duplicate building along the edge of LA City borders. Running QGIS topology checker, we have 18 duplicates out of the 50,000 buildings in the sample data.
You mentioned some issues of nixed buildings along the border, is this related?
There were about 1,000 buildings that intersect with the boundaries of L.A. City. I was able to re-do the difference (in Arcmap, much faster) and I selected only polygons that were fully within the city. So, at worst some people will have missing buildings around the edges. But I don't think it'll be 1,000. It shouldn't be too bad.
@maning it should have been fixed. Hrm. Let me check. I'll see if I can't figure out a solution in the next few hours.
@maning OK. So there are 137 overlaps. Apparently L.A. City decided to draw some buildings outside of its boundary. I didn't anticipate that when I created the last merge. Should I recreate it? Or could we just add a note the import? If you think the latter will be a pain, I can re-kick out the file later tonight.
Should I recreate it? Or could we just add a note the import?
@jschleuss, I think its fine, I can check which blocks are affected and then add a note of these blocks.
I'll take a swing. There may be an issue with the data and buildings that appear to be pre-split right at the boundary. I'll check though.
On Feb 17, 2016, at 6:46 AM, Maning Sambale notifications@github.com wrote:
Should I recreate it? Or could we just add a note the import?
@jschleuss, I think its fine, I can check which blocks are affected and then add a note of these blocks.
—
Reply to this email directly or view it on GitHub.