OpenStreetMap

Improvement of OSM data quality using osmlint and to-fix

Posted by karitotp on 28 September 2016 in English (English)

Improvement of OSM data quality using osmlint and to-fix

To help make it easier for mappers to clean up various quality issues OSM, we have used osmlint to detect dmap ata issues using javascript validators and automatically publish them as tasks daily into To-fix.

In this way mappers are solving thousands of connectivity errors, overlapping ways, unclosed ways, incorrect tags or adding missing tags on the data using satellite imagery daily. Most issues are caused by inexperienced users who may be unaware of how to use the map editor and tags correctly and can be easily fixed by experienced remote mappers. Task descriptions and instructions to fix are provided in the Github repository.

JOSM integration

Fixing issues using To-Fix JOSM plugin

To make it even easier to cycle through a task of interest, experienced mappers can solve issues much faster using the To-Fix JOSM plugin.

Progress of fixes

The community along with the data team at Mapbox has helped cleanup many thousands of issues in the last 2 years.

report To-Fix errors fixed in the last two years

During the course of making these fixes remotely, our team has tried to minimize making incorrect fixes by montioring all changeset comments and taking appropriate steps to improve the algorithm and processing. Now countries with active mapping communities are ignored, changeset comments more descriptive and we recently made a fix to prevent the creation of large changesets that clutter the history view on OSM.

Our workflow to solve to-fix issues

The data team constantly monitors and fixes to-fix errors every day to clean up OpenStreetMap data.

tofix Issues solved in July

A task is assigned to one specific data team member after its completion so that it is constantly monitored for new errors. These new errors are detected and fixed immediately thus reducing the response time to fixing bad edits on the map. Every week, the corresponding tracker on Github is updated with total detected errors and total resolved errors to monitor and document any sudden spikes in issues.

Keeping OSM data error free

The goal is to have a community made map which is constantly growing at the best possible quality. Reaching this goal requires a lot of work to improve the map editors, better validation tools, and the contribution of more experienced mappers. If you are a javascript developer, you can help improve osmlint to catch even more complex data issues or contribute to the development of to-fix, and if you are an experienced mapper you can help fix issues on to-fix.

From the data team at Mapbox

Location: Centro Histórico Ayacucho, Ayacucho, Huamanga, Ayacucho, 05011, Peru

Comment from Nakaner on 6 October 2016 at 17:43

kariototp wrote:

The community along with the data team at Mapbox has helped cleanup many thousands of issues in the last 2 years.

Well, that's not the truth. A few weeks ago I counted all changeset at the changeset planet dump (2016-08-29) which contain the string "#to-fix". I found 363,799 changesets with this string and calculated the number of changesets with this string for each OSM user.

Only 136 users (8 are sockpuppets used for imports) uploaded these changesets. 356,143 (97.0 %) of these changeset have been uploaded by Mapbox employees. I only checked the 84 accounts which had the most to-fix changesets if they are Mapbox employees. The 52 other accounts only have uploaded 100 changesets in total.

Hide this comment

Leave a comment

Parsed with Markdown

  • Headings

    # Heading
    ## Subheading

  • Unordered list

    * First item
    * Second item

  • Ordered list

    1. First item
    2. Second item

  • Link

    [Text](URL)
  • Image

    ![Alt text](URL)

Login to leave a comment