I’ve been digging into the New York Times data APIs recently for a class I’m taking, and the biggest thing I noticed was that there is no way of getting any images past a 75x75 thumbnail. Furthermore, most stories don’t even have images. So I wanted to see if I could both show the stories that the NYT felt were worthy of finding an image for as well as sidestep the 75x75 tiny limitations and create a unique data stream from this.

ImageWorthy does this by
-Connecting the the Times Newswire API
-Extracting the last 200 stories in real time
-Finding stories that include images
-Loading these images into memory
-Analyzing them for average color as well as value
-Creating halftone images superimposed on a background color that is the average color of all the pixels in the image
-Cycling through these every 15 seconds with simple audio effects based on the average value (lightness/darkness) of the image

You can view the project here:

This video is from a desktop version I have running, it looks way more epic projected though without the interface, I need to get footage of that. Also thinking it’d be cool to make this into a screensaver if I ever have a spare second.

I’m using some custom libraries in this that i’m still working on (image filters, loading files from memory, gui element event dispatcher, etc) so I won’t post up all the source because it’ll be a lot of different addons I’ve made that aren’t ready for prime time, but if anyone wants any specific part or explanation let me know for sure!


I like. It’s exciting that people are starting to make data-visualization type projects in OF. Like you said, I bet this looks pretty sweet when it’s large.

Thanks a lot! Yea IDK I have been very interested in data vis lately, its weird that its super popular with processing but not at all with openFrameworks. It might not be super crazy fun or cool to some people but I enjoy it a lot and I’ve been trying to push it lately to openFrameworks stuff. With the tinyXML addon (modified for attributes) I can do all my data parsing in php/MySQL and just import some super slim XML with the core data I need. This way I can also use the data sets with other programs/languages as well if I choose to. I’ve done some stuff parsing within OF/C++ but its pretty tedious/brutal compared to a language like PHP or Python.