inbo / data-publication

🔓 Open biodiversity data publication by the INBO
https://ipt.inbo.be
Creative Commons Attribution 4.0 International
15 stars 7 forks source link

Count data 2018 is missing from summer geese gbif dataset #170

Open SanderDevisscher opened 4 years ago

SanderDevisscher commented 4 years ago

https://www.gbif.org/dataset/2b2bf993-fc91-4d29-ae0b-9940b97e3232

SanderDevisscher commented 4 years ago

The data is present in Recorder => update script should be run

peterdesmet commented 4 years ago

@DimEvil can you tackle this?

DimEvil commented 4 years ago

I don't see the problem. There are counts from year=2018 Last publication: 2019-10-15 Was the import of the data correct? I notice many 0-counts @SanderDevisscher

SanderDevisscher commented 4 years ago

@DimEvil in recorder zitten duidelijk meer dan 44 occurences

SanderDevisscher commented 4 years ago

Was the import of the data correct? I notice many 0-counts

Many 0-counts is to be expected with simultaantelling. The observers get a list of 12 species on which the oberserver needs to pay extra attention. Species not observed in an area creates a 0-count.

SanderDevisscher commented 4 years ago

I checked in recorder and in the INBO - zomerganzen survey there are 280 occurences with at least 1 individual counted during simultaantelling. So somewhere something goes wrong!!

SanderDevisscher commented 4 years ago

Last publication: 2019-10-15

Data of counts were uploaded into recorder the 29th of October.

DimEvil commented 4 years ago

SPATIAL_REF which is published as verbatimLatitude and or verbatimLongitude is not always in the same format recorded in Recorder. This needs to be correct before we can publish the data, because this value is (in Lambert) is used to check if the data was recorded in Belgium or the Netherlands. Simple fix: we also publish the data originating from the netherlands Other fix: make sure SPATIAL_REF is consistent (in Lambert) in the whole dataset.

SanderDevisscher commented 4 years ago

Other fix: make sure SPATIAL_REF is consistent (in Lambert) in the whole dataset.

I checked, they are in the frontend