Flickr Machine Tags – An animal photo example

Similar to the architectural photo that I used as an example in my earlier post, this sea otter photo (taken in the Oceanario in Lisbon, Portugal) has a number of tags included that will help any machine to immediatelly recognize the content of the photo. For this photo I enlisted the help of the tag generator of inaturalist.org.

color:contains=white
color:contains=blue
location:city=Lisbon
location:country=Portugal
taxonomy:kingdom=Animalia
taxonomy:order=Carnivora
taxonomy:family=Mustelidae
taxonomy:genus=Enhydra
taxonomy:binomial=Enhydra lutris
taxonomy:phylum=Chordata
taxonomy:subphylum=Vertebrata
taxonomy:class=Mammalia
taxonomy:subfamily=Lutrinae
taxonomy:species=lutris
taxonomy:common=Lontra-marinha
taxonomy:common=Sea Otter
taxonomy:common=Nutria marina
taxonomy:common=Loutre De Mer
taxonomy:common=Nutria Del Kamtchatka

There are three categories of machine tags that I have used:

  1. Location machine tags
  2. Taxonomy machine tags
  3. Color information machine tags

As you can see most machine tags are from the taxonomy category and show the complete path in the tree of life. Websites that use or have used these kind of tags are the Encyclopedia of Life or the Biodiversity Heritage Library.

Advertisements

Flickr Machine Tags – An architectural example

Having written previously about machine tags on Flickr, here I would like to give an example of used machine tags for a given photo. The photo as can be seen above features the Tokyo Skytree. The Skytree is a relatively modern building and well documented on Wikipedia in terms of metainformation. As such it is pretty easy to give some life to the machine tags on Flickr.

In particular I have used the following machine tags:

location:city=Tokyo
location:country=Japan
dc:identifier=https://en.wikipedia.org/wiki/Tokyo_Skytree
architecture:name=”Tokyo Skytree”
architecture:height=634m
architecture:completed=2012
architecture:architect=”Nikken Sekkei”
abstract:shapes=grid
color:contains=”Skytree White”

First of all I have used a rough geographical framework with the two location tags. Then, as there is an existing wikipedia page, I have used the Dublin Core ontology to make an identifier reference to the wikipedia page.

Thereafter I have given a number of architectural metainformation to the photo, such as name of bilding, height, completion date and architect’s name. Depending on the available information one can add as many tags to this list as possible.

Furthermore I used the tags abstract:shapes and color:contains to give some information on the photo’s composition in general. Please note that due to Flickr’s magic view, identification of the grid pattern as well the colors “white” and “blue” will be done inside of Flickr automatically (and with the right software externally as well). But for those programs that do not analyze the photo itself it is of importance to give that kind of information as well.

An overview to Flickr machine tags and a list of common machine tags

Flickr has enabled machine tagging quite some time ago and there has been a general push in the web towards semantic relations and data structures (e.g. see the archive posts on the Music China Heritage Project). However there seems to not have been a comprehensive “Quick to use” list of common machine tags that photographers and Flickrites can use. In fact while finding a number of high-ranked explanatory posts about the nature of machine tags, I have not found a brief overview of what I (as a typical photographer) can do to enhance my photos.

Updated on 2017-03-17 and 2017-11-06 with further machine tags as found useful as well as additional resources and projects in the web.

General tags applicable to most or all photos:

location:city=* (e.g. Stuttgart, Beijing, New York, etc.)
location:country=* (e.g. Germany, France, etc.)
location:state=* (e.g. Baden-Württemberg, Illinois, Hebei, etc.)
location:street=* (street name)

Every photo is being taken somewhere, and hence the content of the photo has been taken at the LOCATION as specified. Machine tags like these are nice add-ons, while Flickr itself takes the EXIF information to place the photo on the Flickr Map.

wikipedia:en=* (e.g. Paris, Banana, etc.)

This will create a link between the photo and the wikipedia page that is mentioned here. Be aware to NOT enter the whole wikipedia link, but only the title of the page that you are refering to. E.g. a photo representative of Paris should be linked to Paris. But if you have a photo that shows the Eifel Tower it might be better to link that to the Wikipedia article of the Eifel Tower and NOT Paris.

dc:identifier=* (e.g. URL to Wikipedia page or another page about the photo’s content main topic)

The DC Namespace explains the identifier as follows:

An unambiguous reference to the resource within a given context.

and I usually use this machine tag if there is a clear Wikipedia page with further information.

owl:sameas=* (e.g. “Sand Shark”)

I use the owl:sameas machine tag whenever a certain item or animal or person has more than one name and can be identified with another name as well. Animals often have more than one common name and with the owl:sameas tag I do mention each name individually.

Meta data about your camera (even though it is in the EXIF data, some processes do not read those data fields:

camera:maker=Nikon (or Canon, …)
camera:model=D90 (or D70, …)
In principle one could add all EXIF data such as aperture, shutter speed, ISO, focal length, etc. but as a start camera maker and model are sufficient.

For buildings / architecture photos:

architecture:building=* (e.g. Airport / cathedral / church / etc.)
architecture:completed=* (requires a year, e.g. 2015)
architecture:name=* (if a building has a certain name, such as “Schloß Hohenzollern)
architecture:architect=* (name of Architect)
architecture:completed=* (year in which the building as such was completed)

airport:iata=* (if the photo is taken inside an airport or depicts a part of an airport, then you can assign the IATA code here, e.g. FRA for Frankfurt International Airport).

For flower / plant / animal lovers:

taxonomy:kingdom=Animalia (for animals)
taxonomy:kingdom=Plantae (for plants)
taxonomy:binomial=”* *” (replace the stars with the genus and species)
taxonomy:order=*
taxonomy:family=*
taxonomy:genus=*

Information for individual species can be found at Wikipedia. The Flickr group EOL has more information on how they use machine tags.

Furthermore, as pointed out by the BLL there is a great tool that auto-generates the tags for various animals: Taxonomic Tags for Flickr by iNaturalist

For food / dishes / restaurant shots

food:taste=*
food:ingredient=*
food:cuisine=* (Cantonese, Chinese, etc.)

For things in general (e.g. archeological items from museums)

See the separate post about things.

Here are a few examples of machine tags for things:

thing:material=Marble (e.g. Stone, Cotton, Cloth, Steel)
thing:eraoforigin=420-400BC
thing:acquiredin=1857 (Year only)
thing:placeoforigin=Tarquinia (Area of origin)
thing:category=* (e.g. Statue, Figurine, Tool, Weapon, etc.)
thing:length=* (e.g. “35m” or “50cm”)
thing:width=*
thing:heigth=*

If it is known from where an archeological artifact has been excavated one can add a reference to the Pleiades gazetter.

pleiades:place=* (with the identifier as explained here)

For machines and factory plants

Similar to “things” above (which is more general), there are a number of possible machine tags for “machines” / “engines” / etc.

machine:brand=Rüti
machine:yearofcreation=1925
machine:model=* (Model number / type name / etc.)

For Music / Band Promo Shots / Events such as concerts/festivals:

event:type=* (e.g. concert, festival, fair, musical, opera, etc.)
event:venue=* (requires a specific name, such as CGGB or House of Blues Chicago)

music:artist=* (Name of the artist. Be aware of the differences between ARTIST, MUSICIAN and ROLE)
music:band=* (Name of the band, such as Slipknot, Metallica, etc.)

musicbrainz:artist=
musicbrainz:release=
musicbrainz:track=
musicbrainz:label=<MBID

Musicbrainz, one of the large music directories online, has assigned a specific ID for artists, CDs, tracks and labels. On their explanatory page they offer a ay to identify this ID on their website.

foursquare:venue=* (see here for explanation)

Colors / Concepts:

For photos that are more minimalistic in their approach one has the option to include descriptive tags such as:

abstract:shapes=* (e.g. dot / circle / grid / lines)
color:contains=* (e.g. blue / green / yellow / etc.)

Summary and Updates:

  • This should be a handy guide for Flickrites to tag one’s own photos with appropriate machine tags.
  • If you have further tags that are useful, please add them as comments and I will include them in the list.
  • Intitial version of this post as posted on 29.09.2015

Additional resources:

Examples