Home > Public Domain Data > VGI Data Sources: Assessing Completeness and Correctness

VGI Data Sources: Assessing Completeness and Correctness

In a recent article published in the ISPRS International Journal of Geo-Information, Quality Evaluation of VGI Using Authoritative Data—A Comparison with Land Use Data in Southern Germany, the authors investigated some of the concerns regarding data quality and data usability often levelled at Volunteered Geographic Information (VGI) data sources.

The objective of the study, based in the Rhine-Neckar region of southern Germany, was to compare OSM data to the authoritative land use and land cover (LULC) data set ATKIS Base DLM version 6.0. published by the LGL mapping agency (Baden-Württemberg State Office for Geoinformation and State development).

The results for the OSM data completeness and correctness comparison were variable across the different classes of land use in the study area. However some general trends emerged including:

  • Areas with a high percentage of forest cover were the areas with the highest level of completeness and correctness.
  • Other classes (incl. farmland and urban areas) had low levels of completeness but higher levels of correctness; features present were mapped accurately but some features were missing.
  • Other areas (incl. quarry and lakes) had high levels of completeness (most features mapped) but had a greater percentage of incorrectly mapped features.
  • There was a marked difference between rural and urban areas; the study identified higher OSM coverage and thematic accuracy in densely populated areas (more people available/interested in collecting the data?).
  • Some land use classes demonstrated both high levels of completeness and correctness, suggesting they had been mapped for a specific purpose.

Although not intended as a definitive statement of OSM data quality, the study suggested that if full coverage and accurate LULC data was a requirement for a project, then OSM data (at present) may not be the best option. However for certain land use classes, where the LULC information was available it was mostly correct so depending on project requirements OSM data may be a suitable alternative.

As we’ve said many times before on Spatial Reserves, it is not whether the data are good, but rather if they are good enough to meet your requirements.

Dorn, H.,Törnros, T. and Zipf, A. (2015). Quality Evaluation of VGI Using Authoritative Data—A Comparison with Land Use Data in Southern Germany. ISPRS Int. J. Geo-Inf. 4, pp. 1657 – 1670

  1. azolnai57
    November 6, 2015 at 9:04 am

    Great points. Let me take that one step further based on my expereince w UK Ordnance Survey (OS). Meaning, how about using public input ro high grade guv data repositories, in some organised fashion of course (Ordnung muss sein)?
    When OS released admin subdivision polygons (counties, parishes etc.) In 2011, not only did I add some geo. history data for East Anglia from an agro. history book, but I also did some online validation for OS polygons and my attributes via 1SpatialOnline (as part of the then Socium beta program). I turned the error shape files over to OS after meeting them @ London geomob 2012, and a year later they added corrections, which I retested last year:
    To be fair the spotted errors fell below their threshold limit for countrywide datasets, but it illustrated how public guv interaction can work.

    • Jill Clark
      November 6, 2015 at 9:40 am

      Thanks Andrew, a good illustration of the benefits of collaboration. As Anders Pedersen (Open Knowledge Foundation) once noted, more eyes, better data.

  1. November 4, 2015 at 7:39 pm

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: