Harald

If you think of a com­put­er, you imag­ine a stand-alone machine in front of you. In this team, how­ev­er, the com­put­er was only a ter­mi­nal to the cloud. And the cloud is where the data is and the pro­cess­ing takes place.

With the rise of big data and exten­sive pro­cess­ing as the basis of machine learn­ing, there can be no oth­er con­cept than cloud com­put­ing. And we believe, this is also the future of com­put­ing in art history.

The cloud means, you can con­nect to data and services. 

Below you find a sim­ple demon­stra­tion of the con­cept of APIs work­ing together:

  1. The input from the field on the web­site is sent to a web­hook at Inte­gro­mat. That is noth­ing but a URL with vari­ables that are sent to the receiv­ing point.
  2. There, the received key­word is sent to the Rijksmu­se­um API as a request.
  3. It returns a JSON list. In case this con­tains an art­work with an image, it con­tin­ues. Else, it aborts.
  4. From the first entry the image URL is sent to the Google Cloud Vision API that returns labels to the image.
  5. Then, the title of the art­work and the first label as a hash­tag and the image is sent to Twit­ter.

Please enter a key­word and if there is a result in the col­lec­tion of the Rijksmu­se­um, it appears on Twit­ter.


    Mind that not every key­word gains a result and, thus, no tweet is cre­at­ed. This bot is for demon­stra­tion pur­pose only and might be itself ephemer­al.


    This exam­ple shows that with a rather sim­ple con­nec­tion of four end­points, a tweet bot can be realized.

    Has this been dif­fi­cult to achieve? 

    1. We did­n’t need to learn or write any code. We decid­ed to opt for a no-code plat­form. That allows an easy to under­stand approach. 
    2. Even if you write no code, you need to under­stand the basic prin­ci­ples of API end­points, keys, JSON data etc. And you need to think into the log­ic of computers. 

    Dig­i­tal Art His­to­ry does not mean that we will become com­put­er sci­en­tists, but rather that we can talk to com­put­er sci­en­tists about what we need as art his­to­ri­ans. And that requires com­pu­ta­tion­al think­ing. That, we have cer­tain­ly expe­ri­enced in these days.