BUGs: Beautifully Unintended Graphics on Justin's Homepage

BUG Number #7 "portal"

info@justinnk.net (Justin Noah Kreikemeyer) — Sat, 04 Apr 2026 18:40:31 +0200

What happened?

This BUG is the result of me trying to plot an integral curve/trajectory for the ODE system x' = -y; y' = x. Normally, the integral curves of this system would form circles in the x-y plane (fun fact: you can get spirals by subtracting, e.g., 0.5y from y'). However, in this case the resolution of the “measurements” is very coarse. When connecting the measurement points, we get differently rotated pentagons instead of a smooth circle, resulting in this pretty neat looking “portal”. The intersections of these lines even form an “inner circle”. If we were to sample enough points, we would however get the “outer circle” as the correct result. There is Python code available for you to play with.

BUG Number #6 "tsnake"

info@justinnk.net (Justin Noah Kreikemeyer) — Sat, 07 Feb 2026 11:16:30 +0100

What happened?

This BUG happened to my colleague Bjarne when he was playing around with the dimension reduction technique t-SNE. He even took a minute to collect a minimal abstract version of the code that produced it, so you can play around yourself!

By the way, Bjarne writes some great tutorials for the Data Science Club at our institute, go have a look! The most recent one on MCMC I found particularly interesting.

BUG Number #5 "fan"

info@justinnk.net (Justin Noah Kreikemeyer) — Sun, 11 Jan 2026 18:59:27 +0100

What happened?

I sadly cannot remember how this BUG came to be. It seems there was a problem with the shape of the dataset I put into the line-plotting routine, resulting in many more lines than anticipated.

BUG Number #4 "episthetic"

info@justinnk.net (Justin Noah Kreikemeyer) — Sun, 11 Jan 2026 18:52:24 +0100

What happened?

This BUG is the result of vizualising the infected population in an SIR compartmental model. I simulated the model using Gillespie’s stochastic simulation algorithm and plotted the distribution over the amount of infected people at a couple of time points as a violin plot. Turns out I was a bit too generous with the number of discrete time points I choose, resulting in this rather aesthetic abstract image.

BUG Number #2 "termarray2"

info@justinnk.net (Justin Noah Kreikemeyer) — Fri, 09 Jan 2026 23:18:51 +0100

What happened?

This is a second instance of the classic array printing BUG, resulting in another interesting pattern.

BUG Number #3 "quiver"

info@justinnk.net (Justin Noah Kreikemeyer) — Fri, 09 Jan 2026 23:15:04 +0100

What happened?

This is a quiver plot. I plotted the derivative space of a two-dimensional dynamical system, i.e., for each possible state, there is an arrow indicating the tendency of the system if in that state. As we cannot plot all arrows (the state is a real vector), they are plotted on a regular grid. In this case, I made the grid too tight and it turned out to be quite “attractive” (dynamical system pun intended).

BUG Number #1 "termarray"

info@justinnk.net (Justin Noah Kreikemeyer) — Fri, 09 Jan 2026 23:09:56 +0100

What happened?

This is an instance of a classical category of BUG that occurs when you print a long array as a single line inside your terminal or terminal emulator. The wrapping between lines results in regular shifts and interesting patterns emerge. I cannot remember the context of this BUG.