New Yorker Covers: Static
- This project is a blend of my interest in data visualizations and The New Yorker. I had been reading The New Yorker for a few years and loved seeing the variety of art that graced the cover each week. I wanted to see how The New Yorker covers had changed over time. To do this, I collected every cover of The New Yorker from 1925 to 2015, found the average color values, and presented them as one image.
- The images were collected using a PhantomJS script that took a little snapshot of each cover. These snapshots were then sent to Martin Krzywinski’s Image Color Summarizer, which calculated the top ten average color values within the image. Because of the high clustering and precision, each image took about a minute to analyze. To automate the process, I used a Raspberry Pi to run a Python script that collected the average color values and percentages. Even with this setup, the dataset took over 80 hours to complete.
- Once the data was collected, I used Processing to feed in the dataset as a CSV, produce color percentage bars for every cover, and save the result as a PDF. From there I used InDesign and Photoshop to create a layout for the visualization that provided some context as well as a style that was reminiscent of the covers themselves.
- The finished product shows 4,620 covers of The New Yorker on a macroscopic scale, spanning nine decades. It is interesting to see the fluctuations in the color palettes such as the light covers of the late seventies and early eighties, or the dark stretch of covers in the nineties.
Soon after this project was finished, I worked on making the visualization interactive, allowing users to look at specific date ranges as well as the covers themselves. I encourage you to check out the interactive version to explore the covers in more detail.