Skip to content

Wikidata uploading (Feast day project)

Anna edited this page Jun 20, 2024 · 6 revisions

It is very simple to create individual items (Qs) in Wikidata--a log-in is not even required. https://www.wikidata.org/wiki/Special:NewItem allows you to create new items! There are good tutorials for this on Wikidata itself: https://www.wikidata.org/wiki/Wikidata:Tours It will be useful to be an editor, because this is necessary for creating bulk uploads! (In addition to being useful for the usual reasons of provenance history etc.)

Below is documentation about:

  1. Some pages created for this project
  2. Information about properties
  3. Bulk Upload Process
The goal of this project is to create Feast Day Q-entities corresponding to all the feasts existing in the Cantus Database. When these exist, these will be broadly applicable for many musical items, not just chants; for example, https://www.wikidata.org/wiki/Q2522109, which has been flagged as "Created for" (P9883) the Feast of Saint Cecilia.

Some feasts exist on Wikidata already, though in a disorganized fashion. All Saints (https://www.wikidata.org/wiki/Q587), for example, is (as of time of writing) an instance of a Solemnity and a subclass of "religious and cultural festive day," "public holiday," and "anniversary", whereas St. Blaise's Day (https://www.wikidata.org/wiki/Q72900) is an "instance of" a holiday, "religious and cultural festive day," and "Slavic Folk holiday." Corpus Christi https://www.wikidata.org/wiki/Q152395 is an instance of both a "religious and cultural festive day" and a "Solemnity" but a subclass of "Christian holy day". With luck the process of reconciliation will find and correct some of these aberrant items!

The Feast of Saint Cecilia (Q118849459) was created for the purpose of this project. It has the following statements:

  • instance of (P31) religious and cultural festive day (Q375011) of (P642) Christianity (Q5043)
  • instance of (P31) Christian holy day (Q60075825)
  • day in year for periodic occurrence (P837) November 22 (Q3025) (with two references)

For this feast it was possible to find Library of Congress and FAST ID (and somebody has since added a National Library of Israel ID), though this may not be possible for more niche feasts.

See also the Feast of Saint Agnes, https://www.wikidata.org/wiki/Q118842681. The same properties apply, though in this case "named after" and "commemorates" Agnes of Rome (Q210096) were also included (one of these may be more appropriate than the other; TBD).

It is worth noting that Temporale feasts may in some ways be easier; see https://www.wikidata.org/wiki/Q118874088 which defines a day with respect to Pentecost (in this case, in order to correctly label the occurrence of https://www.wikidata.org/wiki/Q1909978)

### Dealing with Properties

Properties are, in general, somewhat more difficult to add, since it has to be discussed and documented at Wikidata:Property proposal first. (cf https://www.wikidata.org/wiki/Help:Properties#Creating_properties): "When after some time there are some supporters, but no or very few opponents, the property is created by a property creator or an administrator (see List of property creators). After creation, relevant documentation is moved to the corresponding property's talk page, where usage of the property is further discussed. When proposing properties, keep in mind that each property should be expected to be used by at least 100 items; if a proposed property cannot be used this many times, it likely should not be added to Wikidata (of course, there are exceptions to this rule)."

Looking through some properties, the approval process for reference numbers in a database tends to be fairly swift, so a property "ID in CantusIndex" should not pose too many difficulties. Others, like "mood" of a piece of music, have been bogged down in discussion for a long time. It may be worthwhile to create properties for CantusID, CantusFeastID, etc, so that these can be uploaded with the list of feasts.

### Uploading the Feasts

Wikidata's Data Donation page (https://www.wikidata.org/wiki/Wikidata:Data_donation) describes some of the process for uploading datasets. https://www.wikidata.org/wiki/Wikidata:Data_Import_Guide also describes the steps:

  1. Choose the data to import (done!)
  2. Create a data set import. This creates the identifier for the data set, like this one: https://www.wikidata.org/wiki/Q125025309
  3. Import the data into a spreadsheet. (We already have many spreadsheets for feasts...)
  4. Define the structure of the data. These are the properties discussed above!
  5. Format the data to be imported. This is a little bit tedious in our case, because the names on Cantus give a Latin feast name ("Michaelis") paired with a translated name of the saint ("Saint Michael the Archangel") which we would like to see as "Feast of Saint Michael". A script would be useful here...(As an aside, we will probably also want to create pages for the saints themselves, in many cases, as well as the date on which they are observed.)
  6. Choose how to import data (we will self import).
  7. Match the data. (The reconciliation step.) Thus far it has proven somewhat tricky for OpenRefine to correctly identify feasts, since they are not ideally formatted in the spreadsheet (see 5.) Trying to match on a "named after" property and treating the saints themselves as people to identify did not seem to help much. (There are also more complicated cases, see below).
  8. Import, using a bot. Within OpenRefine, you can push to Wikidata (as described here:https://www.wikidata.org/wiki/Wikidata:Tools/OpenRefine/Editing/Tutorials/Basic_editing#Uploading_the_changes_to_Wikidata). OpenRefine's instructions (https://openrefine.org/docs/manual/wikibase/overview) note that "to upload your cleaned dataset to Wikidata, you will need an autoconfirmed account, and you must authorize OpenRefine with that account." (An autoconfirmed account is one that has made over 50 edits; adding information about musical instruments is an easy way to get there!) Note that these bots are then attached to the individual Wikidata account; perhaps a project account for all should be made.
### Other considerations:

So far the focus is the "easy" feasts--those named after a single saint, on a single undisputed day. But there are more complicated cases to consider: -Feasts combining multiple people ("Remigius and Germanus") -Feasts with different observances in different places (this can probably be put as different qualifiers in the date property) -Octaves of feasts (or "Sunday in the Octave of XX") -Feasts attached to an event rather than a person sensu strictu: "Translation of the Relics of So-and-so" These will be dealt with once the basic cases are all uploaded!