Random stuff ...
Fourier Transform model for scikit-learn. Can be used standalone. Allows you to predict() future values. In general repetition of the signal may have sufficed, but I decompose the signal to frequencies and then you can plot() or predict() using just top N frequencies, instead the full spectrum.
Linux wc command implemented in Perl.
Mini language, parsers, recursion, lambda, yield ..... in Python.
Check http://ifni.co/articles/programming/recursion, for tutorial.
Simple example of how to scrub html pages in Perl, using Mojo and DOM parsing. You can find the tutorial here : http://ifni.co/articles/programming/web_scrubber
Experimental code for doing Machine Learning, Kalman filters ... etc