UNHCR refugee site mapping Import

From OpenStreetMap Wiki
Jump to navigation Jump to search

[STILL DRAFT]

About

CartONG is a French NGO founded in 2006. Our principal mission is to improve the way aid agencies are using maps and geographic information, and primarily UNHCR who is one of our principal partner. CartONG has been one of the relatively early users of OpenStreetMap in the humanitarian field: we worked with some of the pioneers of the platform and are members of the French OSM association since 2011. We saw the extraordinary potential of OSM in disaster response, and wanted to get more involved so we joined the Missing Maps international project in 2015. Missing Maps (and more generally activities linked to OpenStreetMap) are now an important part of our action. We regularly run mapathons, organize and map projects for our partners, have set up a solid team of volunteer mappers and more importantly validators. We are as careful as possible with data quality and always try to connect with local communities as they endorse to data quality. Promoting open data is in CartONG’s mission statement, given our strong connections with several major actors of humanitarian response, we try to encourage them to use and contribute to OpenStreetMap. We raise awareness on the impact of digital humanitarians and the potential of OpenStreetMap in terms of sustainable development, and advice and support them in opening-up their siloed data. Recently, the UNHCR agree to freed its camp mapping Data Base, and requested support from CartONG to implemented the import.

Goals

Refugee camps are barely mapped and sometime not having a map to navigate and organize aid organization work can be extremely challenging. Over the last years, UNHCR supported by CartONG have been promoting mapping of the refugee camps to get information as recent and accurate as possible on the location of people and infrastructures, to understand their needs and plan the response. CartONG run several mapathons along the year in partnership with the UNHCR in order to map refugee camps in preparation of survey deployment. The UNHCR then deploy field specialist in-country to add attribute and complete the map and trained local staff to update the geographical data base in the long term. Through the years UNHCR has collected a consequent quantity of data, gather together in an internal specific site mapping data base. Those data very useful for UNHCR could also help other aid organization, and refugees themselves. Objective of sharing in the OpenStreetMap platform is to reach a wider audience and ensure the widest possible dissemination of this data.

Schedule

  1. A first pilot to be done on Mauritania camp mapping dataset– December 2018
    1. Data preparation to be done by 22nd of Dec 2018
    2. Communication of import to be done early week of 17 of Dec 2018
    3. Import to be done by end of Dec., if import is supported by the OSM community
  2. Extend import to other countries – starting 2019 (if process is accepted and validated)

Import Data

Data description

UNHCR camp mapping data base gather both wide-ranging data (education, roads, land use, rivers etc) and very specific camp mapping data (health facilities, water distribution, social facilities, etc) for most of the refugee camps under UNHCR supervision.

Background

Data source site: http://maps.unhcr.org/en/apps/campmapping/index.html?site=MRTs004241
Permission: “The provision of data from the Site Mapping database is fully in line with the data sharing dynamic that UNHCR wants to promote. In order to ensure the widest possible dissemination of this data, I am pleased to inform you about the agreement of the UNHCR for putting online those data on OSM platform under ODC-ODBL license under the condition to be cited as owner of the data we provide.”

UNHCR approval

OpenStreetMap License compatibility: verified with the owner

If import is extended to other refugee camps in the next months, the UNHCR will publish an official license approval to use the data in OSM.

Import type:

Mauritania Pilot is a one-time manual import with JOSM (But in preparation of extending import to other countries, semi-automated scripts methods are currently being tested)

Data preparation

Tagging Plans

UNHCR database architecture is quite different compare to the OSM tagging system. CartONG proposed a translated table of OUNHCR data model into OSM tags. The details are available here Datamodel. As mentioned above, UNHCR camp mapping data model contain both general data and camp mapping specific data.

For general data CartONG look for existing OSM tag equivalent:

We used :

  • Map feature and status key/values to assume validity of a tag (approved or defacto).
  • Taginfo, when no status or mapfeature pages where found, to assume the use of a tag, used if largely used.

For camp mapping/humanitarian specific data: there isn’t always an existing OSM tag equivalent. But many tags related to humanitarian and camp mapping have already been proposed, so we try to to gather information from other use case and wiki project such as :

Nevertheless, proposed tags by the humanitarian OSM community hardly go through the all validation tag process. At this stage, we decided to import only the data for which we found a corresponding approved/in use OSM tag In the future, CartONG suggests to open discussion on camp mapping tagging system with the involved actors and go through the process of tag validation, in order to harmonize the tagging system of specific gaps and make OSM data more easily available and searchable.

Changeset tags

Here are the changeset tag we plan to use :

  • comment=UNHCR camp mapping data import, Mbera, Mauritania
  • import=yes
  • url=https://wiki.openstreetmap.org/wiki/UNHCR_refugee_camp_mapping_Import
  • refugee_site=yes (tag in proposal state, used for pilot only)
  • source=UNHCR
  • source:date=yyyy-mm-dd
  • hashtag=#UNHCR;#CartONG;#Mauritania;#name_of_imported_layer (ie worship)

Data transformation

For 2018 Mauritania dataset pilot:

Transformation was done using SQL scripts with : creation of osm table, add geometry, copy data in osm table, replace null value by empty value, update fields following data scheme. Scripts can be shared if asked! It didn't seem relevant to share it here as very specific to UNHCR camp mapping context.


For 2019 extend import:process is still in progress

Data transformation results:

Data was exported in geojson in order to open it in JOSM with validation tool (process in JOSM is detailed in the following sections).

Results can be found here in the shared folder. Link is temporary and will be broken once import is finished.

Data merge workflow

Team Approach:

Import will be undertaken by CartONG OSM team and experienced OSM mappers from our Missing Maps volunteer community, using an import specific OSM user account (CartONG_Imports)

Reference:

This import is referenced in the Import Catalogue. Local OSM community will be contacted and notify of import plans. And as the import is related to humanitarian issues, HOT community will be contacted as an additional measure.

Workflow:

  1. Data transformation, through SQL, tables where converted to fit the OSM scheme, and then exported into geojson
  2. Data is opened into JOSM through geojson plugin, and saved into layer
  3. Data is validated
    1. JOSM validation tool: delete empty fields, and correct remaining errors, manually.
    2. Import osm data in an independant layer, check for OSM existing data to avoid duplicates : as there is only a few data in OSM in the pilot area, and area is local, there is no need to implement automatic checks
  4. Data validated is added to OSM layer through copy / paste at source position (need allow expert mode in JOSM preferences) and then sent to OSM with changeset information detailed in the relative section.

Conflation:

Mauritania Pilot is located within an area with very few POI, conflation is done manually by experienced OSM contributors.

Follow up

Follow up of import on Mauritania pilot is available on wiki page Datamodel and follow up, which details item by item the changeset and data that was imported.