IMPORTING.md

How to import

Getting started

Creating an import account

• OSM best practices require that you do not use your normal OSM account for the imports. Create a new account for this purpose. Usually, it's your existing OSM username followed by _imports (e.g. manings_imports or manings_labuilding). Post your import account username in this ticket.

Getting familiar with JOSM

To contribute to this project, you need to use the JOSM editor. Here are some resources to get you started:

• LearnOSM - http://learnosm.org/en/josm/
• Mapbox Mapping wiki - https://www.mapbox.com/blog/making-the-most-josm/

Check out a task on the tasking manager

• Tasks will be available on http://labuildingsimport.com.
• Priority: we are working on Los Angeles City first, which is broken down by census block groups. Each task performed is one block group within the city boundaries.
• Why? Because different parts of the county have different data problems to watch out for. If we all run into the same problems at the same time, it will be easier for us to help each other and improve the processing scripts and the import workflow.

Import workflow

Download and install the auto-tools plugin

• The good folks at Mapbox created a plugin to merge building shapes sliced by parcel boundaries. You can find it here.

Activating JOSM Remote Control

• Open JOSM and activate JOSM Remote Control. In the JOSM menu, for Windows, select Edit > Preferences... or press F12. For Mac, select JOSM > Preferences... or press, ⌘,.
• Click on the remote control icon.
• Select Enable Remote Control and click OK.

Adding Bing imagery background

• From the Imagery menu, select Bing aerial imagery.

Add L.A. County imagery (for rural areas)

• Bing's imagery outside of major cities can be as old as 2010 (and unusable on the islands). Add L.A. County's aerial tileserver with an offset: From Imagery select Imagery preferences and click on + TMS and add the following URL:

http://cache.gis.lacounty.gov/cache/rest/services/LACounty_Cache/LACounty_Base/MapServer/tile/{zoom}/{y}/{x}

Selecting a task in the Tasking Manager

• Choose which area you want to work on from http://labuildingsimport.com and click Start Mapping.
• Download current data in OSM by clicking Edit in JOSM.

• This will load the existing data from OpenStreetMap (Data Layer 1) and another background layer for the boundaries of the task (Tasking Manager - #2). You will work only within the task boundary.

• Get the .osm file you will import by clicking the link in the Extra Instructions. This will download a new layer in JOSM. At least two layers should be in JOSM: one with the imported data (buildings-addresses####.osm), one with current OSM data (Data Layer 1).

Reviewing the data before uploading

• Review both data layers for possible conflicts.

• Examine tags in both data sets to see if there are any conflicts.

• If there are any problems you don't know how to deal with, do not proceed. Instead, flag the .osm file for a more advanced user to look at. (Use github issues to flag concerns, and/or create OSM notes). Then unlock your task on the tasking manager and pick a new area to work on.

• Preserve the work of previous mappers wherever possible. If existing buildings in OSM are of higher quality:

  • Copy the tags from the import layer version.
  • Delete the building from the import layer.
  • Switch to the OSM layer.
  • Select the building, paste the tags.

• If the imported data are of higher quality, select both buildings and use the Replace geometry tool.

• Run JOSM Validator, and if there are errors, fix them.

Finally, upload it

• Select both layers by shift-clicking them both.

• Right-click the layers and click Merge.

• Merge onto the buildings-...osm layer.

• Click the Upload button, the green up arrow button.

• If you see a "Suspicious data found" warning, click "Continue upload"

• Use the changeset comment: LA County Building Import #labuildings https://wiki.openstreetmap.org/wiki/Los_angeles,_California/Buildings_Import and source: LA County GIS, http://egis3.lacounty.gov/dataportal/.

If you see an Authorization window asking you to log in to OpenStreetMap, log in and remember to use your _import username.

• Go back to the Tasking Manager and click Mark task as done. Another mapper will validate your edits.

What to watch out for

Validate the import with your eyes before uploading!

• Run the JOSM validator. Check for any errors it detects.
• Check for small building parts that should be joined to the main building. We've already found a few examples of these in the data (see issue #19), so make sure you keep an eye out for these. To join small parts, select both polygons and select Tools > Join overlapping Area.

If it's a small sliver, it makes sense that the "proper" data is on the larger object. Delete all the tags on the sliver and join the two objects.

If it's larger, like a strip mall split into pieces, then do:

• lacounty:ain -> ALL
• lacount:bld_id -> ALL
• start_date -> None if multiple or the one option if there's only one
• height -> largest number
• ele -> largest number
• building:units -> none if different

• Inspect everything else with a critical eye! Don't trust that the validator or FIXME tags will catch everything. There may be other bugs that only you can detect. Use your human smarts!

Conflating with existing data

• Look for any overlaps with existing buildings. Existing buildings in OSM are probably more up-to-date than our imported data. Do not assume the imported data is better, mostly likely it is worse!
• Carefully combine the import data with the existing OSM data. If you aren't sure about some tags, ask someone! Especially ask the original mapper!

How to ground-truth the data

• Up-to-date aerial imagery... try several sources.
• See if the street is on Mapillary.
• Go out and check it out yourself! Take a field trip!
• DO NOT USE GOOGLE MAPS OR GOOGLE STREET VIEW.

Identifying new buildings with imported and existing data

• The imagery from Bing was mostly from Los Angeles Region Image Acquisition Consortium (LARIAC). This is the same imagery used as a reference for tracing the imported data. In some areas, the imagery is more updated than the imported data and we can use it to identify newer buildings not found in the imported and existing data.
• Newer building should be identified and flagged. Tag the building with fixme and add a note: Appears in satelite imagery.
• Likewise, buildings maybe demolished and does not exist anymore, delete them.

Communicate communicate communicate!

How to ask for help

• Create issues on this github repo.
• Ask questions on the gitter channel.
• Contact @mappingmashups, @theworksla, @gaufre, @maningsambale.

How to share your progress

• Make sure you close your task on the tasking manager.

How to communicate with other mappers

• JOSM GeoChat feature.
• Twitter hashtag #labuildings.
• Befriend other mappers on openstreetmap.com.