Organised Editing/Activities/Consolidated mapping projects - base roads check vs open data references

From OpenStreetMap Wiki
Jump to navigation Jump to search

Consolidated mapping projects - base roads check vs open data references

Rationale

Rationale, context, involved countries and so on are documented on the dedicated Wiki page: Consolidated mapping projects - base roads check vs open data references

Contact

You can contact alternatively:

Hashtag

The hashtag #BaseRoadsVsOpenData comes automatically with the edits. Prior to April 6th 2019 it was #BaseRoadsCheckProject.

Timeframe

The kind of projects began in January 2015 with Central African Republic. Since it has been extended to at least 9 countries and goes on without ending date.

Tools and Data Sources

JOSM is the only used editor, along with standard aerial imagery (imagery source can change over regions and time) and 500m wide GPX buffers derived from Open Data references which help to identify areas where a road should exist. The Open Data reference is only used to focus on some areas but is never imported: main source for mapping is standard aerial imagery.

In 2018, this methodology as been updated by introducing a priorization tool on mapping road tasks based on some difference between OSM and the reference data, highlighting tasks where roads seems to be missing. This tool provides an uMap that displays data that is automatically updated by osmTrax tool.

Open Data references and their licences, tagging scheme, general mapping instructions and osmTrax tool are documented on the dedicated Wiki page: Consolidated mapping projects - base roads check vs open data references.

Participants

There is not a defined list of participants: practically everybody can join through a Tasking Manager instance hosting one of the mapping projects. However since may 2019, a filter has been set for the projects using Tasking Manager v3 : only mappers with at least an intermediate level (more than 250 changesets) can join.

Between 2015 and 2019, most of participants came to these projects through public events and were experienced mappers :

  • events in West Africa gathering some core mappers from involved coutries
  • events in Toulouse, France, gathering people from local OSM community, in coordination with OSM Mali community

Because a public communication on OSM mailing-lists and a presentation at SotM-FR 2019 are planned for May/June 2019, one can expect that new mappers will participate.

However, one can assess that the mapping instructions complexity participates to encourage rather OSM contributors with concrete mapping experience with JOSM. Firstly because in most case you will have to download à GPX file from Tasking Manager and load it into JOSM to know the area of interest, prior to download manually OSM data on this area which is often quite big. Secondly because mapping instructions implies a JOSM preset usage and are quit complex, warning about different specific cases.

Training/Instructions

No specific training planned except when organising public events, see Participants.

Mapping instructions on the dedicated Wiki Page.

Quality assurance

Validator team (please contact us if you are interested in joining the team !):

Independently from mappers own initiatives to check and improve data quality, our team of validators will:

1) Review edits at least one for each contributor

Use OSMCha to:

  • initialize an empty list of trusted users
  • filter OSM changesets by hashtag and identify the contributors who are not in the trusted user list yet
  • for each contributor: review their edits on the last task they have close or contribute to
    • if edits are OK: add the contributor to the trusted user list
    • if not : contact contributor to provide feed-back
  • iterate regularly

2) Look for suspicious changesets

Use OSMCha to filter changesets by hashtag which specific flags (TODO: provide "Reasons for Flagging" list)

Results

Work in progress. We are looking for a way of providing deeper stats than Tasking Manager v3 default stats and centralized stats for all countries.