GBIF integration

Dave Remsen (GBIF)

Nodes Portal Toolkit

GBIF (Global Biodiversity Information Facility) has initiated a community driven project called the ‘Nodes Portal Toolkit’ (NPT) that should enable communities to deploy, maintain, and extend biodiversity data portals. The platform will also allow community development of new modules with extended functionalities. The initial implementation of the NPT will be built around Scratchpads, linking well with developments in ViBRANT platform. A second version will have extended functionalities modelled on the OBIS data portal. COMBER is expected to be integrated with both ViBRANT and the NPT, offering interested parties a ready-made installation file allowing them to set up and deploy their own citizen-science portals without prior technical knowledge or an on-line application form for a site hosted by the ViBRANT consortium.

Cleaning of GBIF data

GBIF have implemented data quality algorithms as illustrated by David Remsen's talk in TDWG 2011. To illustrate what has been achieved, GBIF have improved geo-referencing processes that enable better matching of records to their intended country of origin. The map below shows raw data originating in the United States.

The large set of data appearing in central Asia, for example, have errors in the longitude relating to E-W notation rather than 360 degrees notation. After correction, the data are all located within the USA.

This map demonstrates the effect of data correction. We can now recognise international waters and offshore islands.

The case of hummingbirds (Trochilidae) shows a different kind of problem. Although hummingbirds are restricted to the Americas, the map below shows numerous observations in Europe.

GBIF have assembled over 50 checklists, taxonomic, thematic and regional, which are used to validate the taxonomic assignment of individual observations. In this case, many European bird species were mistakenly placed in the hummingbird family. Removing those data points results in the map below.

Similar processes are being used to link latinate names to common names which makes display of information much more intelligible to the majority of users.