Identifying flags with AI

If you have seen a flag of a Swiss canton, but don’t know which canton it is for, you have found the solution!

This web app allows you to draw an approximation of a flag you’ve seen, then uses a neural network to try and identify it.

Keplerian Elements for Approximate Positions of Minor Planets

NASA has had simple approximations of orbital elements available on their website for years, however those only list the major planets and Pluto.

I needed approximate positions for big asteroids and dwarf planets for a project and couldn’t find them anywhere, so I calculated a linear best fit based on data from NASA Horizons.

The resulting data is available here: Keplerian Elements for Approximate Positions of Minor Planets

A quick note: Many trans-neptunian objects have orbital elements that move around all over the place. The approximations here are valid over the quoted time line (1950-2100) but you cannot extrapolate them into the future.

Solar eclipse simulator

eclipseThis project originally started because I was bored and thought I could make some prettier graphics than the ones at Wikipedia. But as this stuff usually does, it turned out to be more complex than initially thought, and became a nice experiment that taught me quite a bit about practical astronomy, geographical databases and modern web development.

It is actually quite crazy what sorts of variables you have to consider to accurately model solar eclipses. There are the obvious factors like the ellipticity of orbits, but even smaller ones like Earth’s flattened shape, or the fact that the Earth’s rotation is speeding up because of the last Ice Age’s effects play a huge role and will make your calculations be off by kilometers if they aren’t considered.

Luckily NASA was nice enough to publish their tables of the finished calculations so after a bit of reading up on the interpretation of the data it was quite easy to develop this into a nice looking web app.