Repo cleanup

[rjavila]

Looks like we've now settled into a way of doing things:

1. Clean data in data directory.
2. One directory per test where analysis code lives.
3. Results published in a website.
4. More recent analysis done using pandas (thanks to nicely formatted data files).

I propose we clean up the repository a bit.

1. We can definitely get rid of the tex directory.
2. Might be able to get rid of the lib directory.
3. We should think about converting some of the Jupyter notebooks into pandas scripts.

Any other suggestions welcome.