https://data.dnb.de/opendata/ currently contains 166.056 Mappings from GND to other vocabularies. We only have 43.396 of these as of 2016 (Project MACS).

Would also solve #40 (but only updates every six month).

The sparklines are a great way to illustrate the growth for a property or perhaps also a mapping.

However, it seems that they are not created anew after changes. (e.g., in the last days, I have added lots of GND and RePEc ids, but the sparklines did not change).

Available from the RDF dump at http://nomisma.org/apis#access or via their SPARQL endpoint.

Committing to a dedicated git repository seems to be the easiest way. For better diffs, sort mappings before commiting!

See https://wiki.dnb.de/display/GND/1.5.+Mapping+GND+-+STW and https://d-nb.info/1038596858/34. Mappings are available as PICA Tc records. Mappings to TheSoz, MeSH, Agrovoc, LCSH, Rameau can be made available by DNB as well.

e.g. P2428 has domain human, so add the total number of humans for comparision.

We have some MSC concordances to be added. MSC is revised every 10 years so it must be tracked which version to map against.

Extracted from http://coli-conc.gbv.de/concordances/wikidata/statlog.csv for each property

Probably best to extract via mc2skos from MARCXML

General KOS types (thesaurus, classification...) are given by BARTOC. These types can be shown like tags at each KOS. A list of checkboxes at the top could allow showing/hiding concordances because the current list becomes too long.

Mit einer neuen Version von wdmapper, bei der die Annotations ausgelassen werden können, sollte es möglich sein, auch große Mappings abzurufen

git grep http://d-nb.info/gnd/ and grep -r http://d-nb.info/gnd/ should help

E.g. http://rvk.uni-regensburg.de/nt/TA-TD should be "TA - TD" instead of "TA-TD".

In particular Places and Times should be easy to map.

More than 1.5 Million mappings but still possible to download on local machine:

$ time wdmapper get P1566 -g '' -o P1566.txt

real	1m27.590s
user	0m52.872s
sys	0m5.252s

On the server, however the call dies after half an hour with segfault or similar. Further debugging required. The SPARQL query is

SELECT ?item ?target WHERE { ?item wdt:P1566 ?target }

Looks like a bug in wdmapper (?)

The current file gnd_ddc_crisscross.csv seems to be truncated (see last line). Current mappings could be extracted via mc2skos from MARC dump.

Due to the number of mappings this can better be done on the server. This includes:

• NSK-BK
• RVK-BK (multiple concordances)

See http://zbw.eu/stw/version/latest/download/about.en.html

Every broader or narrower relation we found that was generated by the CrissCross project seemed to be the wrong way round.
(We searched for example for fruit such as "Kirsche", or "Emmentaler Käse", or "Bengalkatze".)

Available at https://op.europa.eu/en/web/eu-vocabularies/alignments

See https://github.com/jamiebicknell/Sparkline/blob/master/sparkline.php for an implementation

See https://github.com/gbv/jskos-php-examples for access to KOS. This should be extended to mappings, e.g. from Entity Facts, Wikidata.org (live), VIAF.org, and sameAs.org etc.

Several EuroVoc mappings exist, see https://op.europa.eu/de/web/eu-vocabularies/alignments.

The list of EuroVoc concordances can be queried via http://publications.europa.eu/webapi/rdf/sparql:

DEFINE input:same-as "yes"

PREFIX cdm: <http://publications.europa.eu/ontology/cdm#>
PREFIX owl: <http://www.w3.org/2002/07/owl#>

SELECT ?s ?title {
  ?s cdm:work_dataset_keyword "eurovoc alignment" .
  ?s cdm:work_title ?title
}

A "Started at {date} with {size} items" tooltip could add a lot of information.

The bash extension that allows uniq -D (which is used in the Makefile) is apparently not available on macOS and I couldn't find out how to install that. Maybe uniq -d is enough? (that works on macOS)

kos-registry is now included as a git submodule, but there are no instructions that it is necessary to fetch that. So we need something like this:

# Either
git submodule init
git submodule update
# Or
git clone --recurse-submodules https://github.com/gbv/cocoda-mappings.git

Imported mappings from concordances should get an URI to be annotateable. Could be based on mapping content identifier plus partOf (see gbv/jskos#68)

Since moving kos-registry to a submodule, make in the csv folder won't run anymore because file 'kos.yaml' doesn't exist. Either the Makefile has to be adjusted to use the file from ../kos-registry/, or we need a symlink.

wdmapper does not harvest mappings if

• property lacks URL template (P1630): P1691, P1190, P667, P1189, P2347, P1762, P714, P2452, P305
• property has multiple URL templates (P1630): P486, P1256, P1257
• property is not of type external-id or string: P2643

For make to work in the csv directory, some perl dependencies are necessary. Maybe include something like cpanm Catmandu Catmandu::MARC. Also include documentation about what make and make ndjson does (and that it exists in the first place).

Short of filing an issue for sortability, I detected in the code that the table already is sortable - which is great. Two additonal suggestions:

• default sort by KOS name (which perhaps what most people would expect. Another reasonable default could be size descending)
• add some indicator, that the columns are sortable, and which sort is active

The list of concordances should link to dump files as shown at http://coli-conc.gbv.de/concordances/

Use %20 instead of _ and respect spaces in number ranges.

Der Umfang indirekter Konkordanzen wird zwar ermittelt aber noch nicht angezeigt. Sehr schön wäre eine Darstellung mit Venn/Euler-Diagrammen: https://github.com/benfred/venn.js/

Both directions available:

• https://www.questionpoint.org/crs/html/help/en/ask/ask_map_lcctoddc.html
• https://www.questionpoint.org/crs/html/help/en/ask/ask_map_ddctolcc.html

Question: Which license does that data use? Can we get it in a better format than an HTML table?

Also, we don't have the LCC in Cocoda yet: gbv/jskos-data#21

To align with gbv/jskos-server#16:

• fromNotation (instead of sourcenotation)
• toNotation (instead of targetnotation)
• type
• ...

See also https://github.com/gbv/cocoda-mappings/blob/master/csv/README.md

Duplicates are included in *.err files