pascalheid / thesis Goto Github PK

In the folder latex you can find my tex filex which compile the pdf which I handed in as my final thesis. It relies on the figures in the figures folder. These as well as all other figures and tables of my thesis can be recreated by running the replication notebook. This notebook relies on several python modules from which I import some functions. These modules can be found in the python folder. The data contains all the stored results from my simulations as well as the data from Iskhakov et al. (2016).

I set up travis to execute my notebook as a replication check and the CI service of github to compile my thesis as pdf. My replication can also be accessed via mybinder. All necessary badges are included at the very top.

In my thesis, I heavily rely on the ruspy package for which I made some contributions in the context of my thesis. These contributions can be seen in the Pull Requests #42 and #46. Additionally, I needed an implementation of the BHHH algorithm in Python which I had previously done in a project for Effective Programming Practices for Economists. During my thesis, I revised the BHHH and added a PR to estimagic.

The current state is that I am still awaiting the PR for estimagic and the second PR for ruspy to be merged. Due to that, the travis build currently excludes running one block of code which replicates the true demand level on page 29 of my thesis (as the contributions are not available via conda, yet). This also means that if you want to run my whole notebook including the lengthy simulations, you currently have to - additionally to installing my provided environment - download the source code of estimagic and ruspy from my two PRs and install them in editable mode. If you do so, my whole replication notebook runs through.