This should probably go into a funstuff section somewhere, but I used some moments of free time today to upload a script I have written some time ago to github. It’s a very simple piece of code that grabs images tagged with a specified word and, by looking at which tags appear together, creates a co-tag graph file in .gdf format. You can get it from here or run it here. To test how it scales – and to finally know what teens (apparently tumblr’s main audience) dream of – I tried it with 500 sets of 20 images for the tag “dream”. This leads to some 7K distinct tags and after some filtering, that’s what comes out (click image for lager view):
dreams
Node size is occurrence count and color (blue => yellow => red) is betweenness centrality. Apparently, love is still a thing out there. Nice.

This may become an actual tool further down the road, but maybe it’s already useful to somebody as is.

EDIT: Try it out here: https://lab.digitalmethods.net/~brieder/tumblr/tagnet/

Post filed under folksonomy, social networks, softwareproject, tumblr, visualization.

3 Comments

  1. Please map what people above 30 + dream of !!!

    • Hi, unfortunately, I don’t have access to demographic information, I simply said “teens” because that seems to be the main demographic of tumblr…

  2. Ha, ha… Nice piece! Made me laugh…

Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>