The Paradise Papers represent one of the largest public data leaks comprising 13.4 million confidential electronic documents. A dominant theory presented by Neal (2014) and Griffith et al.(2014) concerns the use of these offshore services in the relocation of intellectual property for the purposes of compliance, privacy and tax avoidance. Building on the work of Fernandez (2011), Billio et al.(2016) and Kou et al.(2018) in Spatial Arbitrage Pricing Theory (s-APT) and work by Kelly et al.(2013), Ahern (2013), Herskovic (2018) and Proch ́azkov ́a (2020) on the impacts of network centrality on firm pricing, we use the market response, discussed in O’Donovan et al.(2019), to characterise the role of offshore services in securities pricing and the transmission of price risk.
This is your new Kedro project, which was generated using Kedro 0.15.6 by running:
kedro new
Take a look at the documentation to get started.
Dependencies should be declared in src/requirements.txt for pip installation and src/environment.yml for conda installation.
To install them, run:
kedro install
and either:
conda install --file src/conda.yml
or:
mamba install --file src/conda.yml
For those using mamba alongside the Anaconda package manager
For those looking to run the pipeline, a docker container as been provided for use in reproducing our analysis. This can be run to recreate our environment using:
docker-compose run python
This will assume all raw data sources have been downloaded
You can run your Kedro project with:
kedro run
This will assume a IEXCloud api key is provided in config/local/secrets.yml as:
iex: YOUR_KEY
For researchers using our downloaded data, they may use:
kedro run --tag=local --parallel
This will assume all data is already available in the data directory.
The pipeline currently follows the current io topology:
Our analysis does provide py:percent format notebooks which provides discussion over our exploratory work. This can be run either directly in a notebook environment that support py:percent, like VSCode, or may be converted to ipynb files using:
jupytext --to ipynb *.py
In order to use notebooks in your Kedro project, you need to install Jupyter, this has already been included in our conda environment but is an important flag for those running our exploratory analysis:
pip install jupyter
For using Jupyter Lab, you need to install it:
pip install jupyterlab
After installing Jupyter, you can start a local notebook server:
kedro jupyter notebook
You can also start Jupyter Lab:
kedro jupyter lab
And if you want to run an IPython session:
kedro ipython
Running Jupyter or IPython this way provides the following variables in
scope: proj_dir, proj_name, conf, io, parameters and startup_error. This abstracts and simplies interaction with our versioned datasets or API endpoints and ensure data is being loaded and serialized in a consistent manner.
Project documentation has been provided to guide users through our code. This can be found at the docs/index.html.
To build API docs for your code using Sphinx, run:
kedro build-docs
See your documentation by opening docs/build/html/index.html.
A precommit hook has been added to this project to automatically build this documentation to be served on Github Pages. This should allow users a familiar interface for navigating the documentation accoumpanying our code.