Commons:Flickypedia/Data Modeling

• Main page
• About the tool
• Community engagement
• Data modeling
• FAQ
• Project Plan
• Team

Modeling the Data Connection

A document to outline Flickypedia’s alignment with Structured Data on Commons

Hi. This is work in progress. If you have comments, do please add to the Discussion tab!

For any given image uploaded using Flickypedia or backfilled with Flickypedia, we automatically map the following data:

Always mapped

‎Flickr photo ID (P12120)	This is meant to be an unambiguous, canonical property for finding existing copies of a Flickr photo on WMC. It avoids all the hassle of dealing with different varieties of Flickr URL. For more info/examples, see the property proposal. Example: P12120 → 326048874
creator (P170)	This is the Flickr photographer, i.e. the user who uploaded the photo to Flickr. Otherwise, we create a creator (P170) with qualifiers: Flickr user ID → Flickr user ID (P3267) Author name → author name string (P2093) URL → URL (P2699) Example: without Wikidata entity P170 → P3267: 54289096@N00 P2093: zen Sutherland P2699: https://www.flickr.com/people/zen/
source of file (P7482)	We link to the original photo on Flickr using source of file (P7482) → file available on the internet (Q74228490) with qualifiers: Described at → Flickr Photo page URL → described at URL (P973) Operator → Flickr.com → Flickr (Q103204) URL (to full-size image) → URL (P2699) Example: P7482 → Q74228490 P973: https://www.flickr.com/photos/merula/7226882/ P137: Q103204 P2699: https://live.staticflickr.com/5/7226882_1b236da680_o.jpg
copyright license (P275)	We copy the currently-stated license from Flickr. Example: CC BY 2.0 P275 → Q19125117
copyright status (P6216)	We populate based on the currently-stated license on Flickr. Example: CC BY-SA 2.0 = copyrighted P6216 → Q50423863
published in (P1433)	We save the date a photo was posted to Flickr in published in (P1433) → Flickr (Q103204) with a qualifier: publication date (P577) → date the photo was posted. Example: P1433 → Q103204 P577: 17 September 2019

Sometimes mapped

published in (P1433)

We save the date the photo was taken in inception (P571). Not all photos on Flickr have a public "date taken" value, e.g. if the uploader doesn’t know when the photo was taken. In this case, we won’t save a date taken on WMC. Photos on Flickr support different levels of granularity, which we map to precision in the Wikidata date model Date taken on Flickr can be marked as "circa"; if so we add qualifier sourcing circumstances (P1480) → circa (Q5727902) Examples: taken 18 April 2008 P571 → 18 April 2008 taken circa 1930 P571 → 1930 P1480: Q5727902

coordinates of the point of view (P1259)

We save the location the photo was taken in coordinates of the point of view (P1259). Unlike Wikimedia, Flickr doesn't distinguish between the location of the camera and the subject. We chose P1259 as the best match for how Flickr photographers use the field. Not all photos on Flickr have location information, e.g. if the photographer has made it private. If there isn't public location data, we won't save a location on WMC. i.e. we'll only copy across location information that you could see by visiting the photographer's page on Flickr – we’re using the location information returned by Flickr’s public API. Example: P1259 → 56°22'34"N, 62°9'10"E

Never mapped

Flickr title and description	We’ve decided not to bring these across – this text may not be appropriate for Wikimedia Commons (e.g. more informal/personal, not descriptive), and we may not be able to license this text as CC0.
Flickr tags	Flickr tags are more folksonomic that the structured data definitions used in Commons, e.g. they're not as specific as Depicts. We store them in the Wikitext with links back to Flickr, but not in the SDC.
EXIF metadata, e.g. captured with	There are already bots that fill in these properties with data from Flickr, and we didn't see a need to duplicate their work.