I wanted to give an explanation of my header image. It is based on the birthday problem, something I plan to introduce in the course I am developing for this August. I don’t generally just make art based on a statistical concept that I plan on teaching, but Brown’s division of Biology and Medicine (where I am a doctoral candidate in Biostatistics) is going to have an exhibition on “The Art of Science” in the Fall, and I got inspired to put something together.
Here is the premise of the birthday problem: ask someone how many people would have to be at a party in order for there to be a 50-50 chance that at least two people have the same birthday. If the person you are posing this question to hasn’t heard of this problem before, their intuition might be that it would be half of the number of days in the year, about 182. But (under certain simplifying assumptions) it is actually only 23. This can blow peoples minds, which is why it is so often used in introductory Statistics courses.
The header image, which I have yet to name, is based on this birthday problem. I used R to generate 23 numbers with replacement, which could range from 1 to 365, to represent the birthdays belonging to 23 people at a party. I simulated 40 parties of 23 people. If the party did not have two people with the same birthday I made a white horizontal band and denoted peoples’ birthdays with a small blue rectangle. If someone in the 4th party had a birthday on the 85th day of the year, this would be represented in the image with a rectangle at the coordinates (x =85, y=4). If the party had at least two people with the same birthday, I made a yellow horizontal band and denoted people who did not have a duplicated birthday with a red rectangular box, and people who did have a duplicated birthday with a black rectangular box. I decided to highlight any date with a duplicated birthday by including a black vertical band on those dates. This gives the image a sense of depth as the black and yellow perpendicular bands appear to be woven together.
I was particularly inspired by Mondrian, with his color palate and his use of intersecting lines.
I also included gray and white bars along the x-axis to represent the months in the year. January, March, May, July, September, and November are in gray, and the others are in white. In the y-axis I have included little gray boxes which let you know the width of the bands that represent each of the 40 parties that were simulated.
I was thinking about making a brief movie clip with different realizations to show that on average half of the horizontal bands are yellow. My hope is that this communicates how common “coincidences” really are in a new way, and that it is aesthetically pleasing. Whether or not this communicates anything, or looks nice, I had a fun time making it, and I’m all for blending statistics and fun.