Today, I learned how to access the Instagram API and ran my Instacollage sketch using Instagram photos pulled from the web. (Many thanks to Uttam who helped me figure out how to work with the Instagram API!)
Less talk, more collage!
First up, “Instacool collages” which combine recent “popular” images (photos with thousands of likes).
Many of the Instacool collages have at least one or more of the following: sports, scantily clad women/cheerleaders and expensive cars/bling. Perhaps not surprising?

I suppose you could try and game the system and throw up images on Instagram of scantily clad women wearing lots of bling and drinking alcohol while riding expensive cars and see how many likes you get!
I also tried combining images with different (or the same) tags. The results were unpredictable and random, with the occasional gem.
Cats + dogs = bound to be awesome?

Or, how about just dogs?

I decided I might get some interesting results using photos tagged with emotions.
How about love + hate?

This is when I ran into combinations with words! The collages suddenly take on a confessional aspect.

How about sad and hate?

Bored and sad?

Lastly, Instagram, what is the meaning of love?

For every 50 collages my sketch created, there were maybe a handful of collages that I found interesting. I think the instacollage idea is stronger when curated as the random generation returns a lot of meaningless results. One idea I have is to create an Instagram “Instacollage” user profile and throw up the best Instacollages as the feed.
Lastly, I am disappointed cupcakes + cats did not produce any good results!