I wanted to see what would happen when entries from Overheard In New York were matched up with related images from Flickr. I thought that if I could extract keywords from the text, I could retrieve public images from Flickr and the results could be hilarious. Or weird. Or nonsensical. Hence the name: nonsequitr.com.
The text on the left side of the screen is from Overheard. I found a widget that they offer on their site, and with some simple Jquery and PHP (I'm still learning), I was able to extract the random quote from the feed in that widget. Once I had that, I used a script called "Keyword Extract" that pulled common words (mostly prepositions, adjectives, adverbs, and even some verbs) out of the quote, and that left me with a pretty decent keyword list.
From there, I used those keywords to search Flickr, and that the image that gets returned is the one you see. Sometimes the pairing makes sense. Other times, not so much. But that's what makes it fun.
Other than sharpening my web development skills (and having a few laughs), the biggest takeaway I got was building the "common words" list. It was eye-openeing to see how computers look at words and sentences versus how a reader would.
Try it out! Nonsequitr.com