Open cgiovando opened 5 years ago
@cgiovando After providing training on the use of TM, iD Editor and JOSM, we devised a workflow for identifying and extracting informal settlement boundaries that we had high confidence in. Jamon the OSU team are writing up our workflow and will share shortly, but in a nutshell the process is as follows:
Please note that some of the GeoJSON files saved to the Google Drive are marked as modified, this means that the original boundary was altered slightly to encompass the expanding area of the refugee camps. This is to provide Mark with the most up to date settlement boundaries, without making changes to the data in OSM at this point.
An initial TM project was created using a selection of known UNHCR refugee camps: https://tasks.hotosm.org/project/5668
The original subset included 254 point sites across all of Africa. After applying a 3km buffer on each point and dissolving the overlapping areas, a total of 83 distinct tasks were created.
Training on TM and OSM mapping tools was conducted at Oregon State University on Jan 15 by @mataharimhairi. Participants were asked to map settlement boundaries according to instructions in the TM project.
Output from this preliminary mapping task will be used as training data for machine learning processing by @wronk
@mataharimhairi @JamonVDH please add any comments, lessons learned and recommendations from this first project.