Photo Mining

    I thought I'd share my final project for an art class I took last spring quarter. This is much different than a lot of my other work, but it was a fun project to work on. Watch the video below and then click "Show More" below for an explanation of the project. (Make sure to watch the video first, though, because the explanation spoils it.)

    I made this set of images for a class called Algorithm and Art - a course that covered everything from programming art in Processing to using 3D printers and laser engravers. A interesting aspect of computational art is the use of random seeds to generate different iterations of pieces that have a similar style. Working from this idea of randomization, I decided to create an Excel spreadsheet that generated everything you need to take a "random" photograph - randomized longitude and latitude, rotation, height, pitch, focal length, and time of day. I imposed constraints on location to keep things within Hyde Park and constraints on the pitch so that I didn't have tons of photos of the sky and ground. All photos were taken between noon and midnight because time doesn't really exist before noon for college students. Finally, I couldn't follow the latitude and longitude directions faithfully if they landed on private property, but I tried to get as close as possible on public property. (My professor thought I should knock on people's doors and explain to them that I needed to take a photo inside their house because my Excel spreadsheet led me there, but I wasn't ballsy enough to do that.)

Instructions for randomized photos. 

     I took the resulting images and tried to tease out themes in the photos. This part was inspired by a statistics class I took earlier in the year where I was running regressions on large data sets. One danger of doing statistical analysis on large data sets is that you can find trends that look like they represent real-life phenomenon but are actually due to chance. This project grew out of that idea of finding fake trends in random data.
     The first section of the video is "Green and Grey - Exploring the Interplay of Vegetation and Cement in Urban Environments" because there were trees and sidewalks or streets in pretty much every photo. I wanted a pretentious artsy-academic sounding title so that people though I was trying hard to be profound. I repeated this process for the second section, called "Cars and Paths - An Intimate Look at our Means of Transportation" because there were cars in nearly every shot. The final section reveals that everything is random, and hopefully it becomes obvious that the themes were fabricated. Were you fooled?