Slik bruker de beste museene åpne data

Å gjøre data tilgjen­gelig er rel­e­vant for de fleste fagom­råder i offentlig sek­tor, ikke bare innen økono­mi eller fiskeri. Nylig har jeg skrevet flere artik­ler om hvor­dan insti­tusjon­er som for­val­ter kul­tur­ar­ven arbei­der aktivt med å åpne sine data og dig­i­talis­erte verk for utviklere og for all­mennheten. Dette gjelder arkiv­er, bib­liotek­er og museer.

En av sak­ene går gjen­nom sta­tus på fel­tet og knyt­ter utviklin­gen til fores­låtte endringer i reg­lene for offentlig sek­tors data. En annen sak tar spe­sielt for seg erfarin­gene det berømte Rijksmu­se­um i Ams­ter­dam har høstet etter at de ved årsskiftet la om til en åpen­het­slin­je. Råd­giv­er Lizzy Jong­ma for­t­alte meg i et epost-inter­vju hvor­dan museet tenker om åpne data og tilgjen­gelig­gjøring av dig­i­talis­erte verk. Neden­for er inter­vjuet i sin helhet.

“Melkepiken” av Jan Ver­meer. (foto: Rijksmuseum)

Et eksem­pel museet bruk­er er Jan Ver­meers “Melkepiken” (The Yel­low Milk­maid) fra ca. 1660. Mester­ver­ket fantes i tusen­vis av kopi­er på net­tet, de fleste dårlig utførte repro­duk­sjon­er som ikke yter orig­i­nalen ret­tfer­dighet. Ved å legge ut sin egen kopi i høy oppløs­ning håper museet å motar­bei­de en slik uheldig utvikling, som har ført til at besøk­ende i museets butikk ikke tror postko­rtene vis­er det rik­tige bildet!

Could you describe how pub­lish­ing a high-res. image with open meta­da­ta helped? What kind of effects has it had?
Lizzy Jong­ma: “Since we start­ed pub­lish­ing our meta­da­ta with a URL to our high/higher res images (the images are jpegs with a file size up to 5mb/ 300 dpi. We have Tiff images in much high­er res­o­lu­tions but we don’t put these images online) we’ve had a lot of atten­tion for our col­lec­tion. From media, from app builders, pro­gram­mers and from our col­leagues. Their main focus was with the fact that we shared our high res images online with a CC-BY license.

It has helped us tremen­dous­ly in pro­mot­ing our (online) col­lec­tion. We want every­body to know (parts of) our col­lec­tion and use our dig­i­tal col­lec­tion for any/your own pur­pos­es: if you want to do a pow­er­point about Ver­meer or Rem­brandt, then please use our images and infor­ma­tion instead of the ugly and bad images that are on the inter­net (that is what we mean with the Yel­low Milk­maid anec­dote. If you use Google images to find this paint­ing of Ver­meer, then you will find very ugly images that don’t do any jus­tice to the beau­ti­ful, fresh orig­i­nal paint­ing). We are a pub­lic insti­tu­tion and it is our mis­sion and goal to share our col­lec­tion and knowl­edge with every­one. We have seen an increase in vis­i­tors to our online col­lec­tion. And we also ben­e­fit from that: we get e‑mails every day from experts in all sorts of fields that help us anno­tate our col­lec­tion or cor­rect us if we’re wrong.

Since we put an API online (for devel­op­ers and pro­gram­mers so that they can inte­grate our col­lec­tion in their own appli­ca­tions or web­site) we’ve had over a dozen new (free) apps built with our col­lec­tion. Some are nice and some are real­ly use­ful and inno­v­a­tive. On www.rijksmuseum.nl/api you can find a list with apps. Arkyves built a great app: an icono­graph­ic brows­er to search our col­lec­tion the­mat­i­cal­ly. The brows­er is avail­able in Eng­lish, Ital­ian and French. Our col­lec­tion meta­da­ta is in Dutch and this is the first inter­na­tion­al entry to our collection!

AB_C Media built a “faces of the Rijksmu­se­um” app: using face recog­ni­tion soft­ware they were able to iso­late all the faces on our paint­ings. This can help us in cre­at­ing more spe­cif­ic, more detailed descrip­tions of our paint­ings (for instance pin­point­ing every­one on the Night­watch).

Our Open Data was also har­vest­ed by nation­al and inter­na­tion­al plat­forms and por­tals: Art­stor, the Dutch muse­ums col­lec­tion web­site, Ken­nis­net (the nation­al infra­struc­ture for e‑learning in schools in the Nether­lands). We are now part­ner­ing with the VU Uni­ver­si­ty (an Ams­ter­dam Based Uni­ver­si­ty with which we have been part­ner­ing for years in devel­op­ing seman­tic tools) to devel­op crowd­sourc­ing and seman­tic anno­ta­tion tools based on our Open Data and so on. In short: we’ve had a lot of atten­tion and con­crete results that help us pro­mote and fur­ther anno­tate our collection.”

Are there now few­er bad copies around on the web? Do more peo­ple find the high-res image on your web­site and is it being repub­lished on oth­er sites?
“No: the bad copies aren’t gone. Unfor­tu­nate­ly. We’ve only had our Open Data online for three months, and unfor­tu­nate­ly most bad things stay on the inter­net for­ev­er. We do see an increase in the usage of our online col­lec­tion and images. By giv­ing Wikipedia, Euro­peana and Ken­nis­net access to our meta­da­ta and images we hope that kids and teach­ers and so on will get to know the good (high res, colour bal­anced, not pho­to­shopped) images. We’ve also done a lot of work in opti­miz­ing our online col­lec­tion for search engines like Google and we hope that our col­lec­tion will end at the top of Google search­es over the next year, but decreas­ing the amount of poor images is a long term strategy.”

In gen­er­al, what is the Rijksmu­se­um’s pol­i­cy regard­ing pub­lish­ing meta­da­ta and mak­ing pub­lic domain works avail­able? How will mak­ing more data and con­tent avail­able help your work?
“The Rijksmu­se­um pol­i­cy is (as said) about shar­ing our col­lec­tions and knowl­edge with every­one (if pos­si­ble: we do have to stay with­in copy­right bound­aries). We now have 100.000 objects in our Open data col­lec­tion (all with images) and 250.000 objects (not all with images) in our online col­lec­tion and we will con­tin­ue to dig­i­tize our col­lec­tions as part of our dai­ly work. Shar­ing helps us to pro­mote our col­lec­tions. The pub­lic wants to see our col­lec­tion, cre­ate new prod­ucts, build new appli­ca­tions (show­case appli­ca­tions with our col­lec­tion) and share knowl­edge with us about the col­lec­tion. Open Data didn’t change our work (we were already dig­i­tiz­ing and anno­tat­ing our col­lec­tions) and it helps us ful­fill­ing our mis­sion and goals.

We now advice oth­er Dutch muse­ums (togeth­er with Ken­nis­land, the CC rep­re­sen­ta­tive in The Nether­lands) in imple­ment­ing Open Data and Copy­right issues con­cern­ing open data. We also entered our API in dif­fer­ent Hack Bat­tles and will present our project at Muse­ums and the Web.”