topics_causal_inf

Topics Course Causal Inference Summer 2024

https://github.com/buddejul/topics_causal_inf

Science Score: 31.0%

This score indicates how likely this project is to be science-related based on various indicators:

  • CITATION.cff file
    Found CITATION.cff file
  • codemeta.json file
    Found codemeta.json file
  • .zenodo.json file
  • DOI references
  • Academic publication links
  • Academic email domains
  • Institutional organization owner
  • JOSS paper metadata
  • Scientific vocabulary similarity
    Low similarity (12.2%) to scientific vocabulary
Last synced: 10 months ago · JSON representation ·

Repository

Topics Course Causal Inference Summer 2024

Basic Info
  • Host: GitHub
  • Owner: buddejul
  • License: mit
  • Language: Python
  • Default Branch: main
  • Size: 2.24 MB
Statistics
  • Stars: 0
  • Watchers: 1
  • Forks: 0
  • Open Issues: 0
  • Releases: 0
Created almost 2 years ago · Last pushed almost 2 years ago
Metadata Files
Readme License Citation

README.md

Topics Causal Inference

Replication repository for Causal Inference Topics Course Summer Semester 2024.

The repository follows the src/bld structure. All input (source) files generating outputs are stored in the src folder. All generated output (see below for instructions) will be stored in the bld folder. The project is completely replicable based on the provided repository, with the exemption of the generated WGAN files (details see below).

Instructions

To run the replication, Python has to be installed on the system.

First, after cloning the repository, create the virtual environment by typing

shell mamba env create

in the shell. Note the current path of the shell needs to be the cloned repository.

For building the project, I use a workflow management system called pytask. pytask facilitates reproducible analysis by automatically discovering tasks and evaluating tasks only when the code or dependencies have changed.

To build the project, first activate the virtual environment and then run the build:

shell conda activate topics_causal_inf pytask

However, it is advised to run the tasks collected by pytask in parallel by specifying

```shell pytask -n [n_workers]

Alternative: Starts os.cpu_count() - 1 workers.

pytask -n auto ```

With 11 workers on my private computer, building the full project takes about 90 minutes. To reduce runtime, the simulation settings in config.py might be changed.

In particular,

python N_SIM = 40 N_SPLITS = 10

control the number of total simulation runs as well as the number of splits for the generic ML procedure.

Note that I generated the inputs to the WGAN-based simulations externally on a Google Colab server. The results have to be stored under src\topics_causal_inf\wgan_generated\. Alternatively, DG_TO_RUN = ["standard", "wgan"] in task_wa_replication.py has to be changed (i.e. remove "wgan").

Note: Sometimes the compilation of the LaTeX document fails on the first try, but rerunning pytask then usually works. Else, this step can just be ignored.

Owner

  • Name: Julian Budde
  • Login: buddejul
  • Kind: user
  • Location: Bonn, Germany

Econ PhD, Uni Bonn

Citation (CITATION)

@Unpublished{budde2024,
    Title  = "Topics Causal Inference",
    Author = "Julian Budde",
    Year   = "2024",
    Url    = "https://github.com/buddejul/topics_causal_inf",
}

GitHub Events

Total
Last Year

Dependencies

.github/workflows/main.yml actions
  • actions/checkout v4 composite
  • codecov/codecov-action v3 composite
  • mamba-org/setup-micromamba v1 composite
  • r-lib/actions/setup-tinytex v2 composite
pyproject.toml pypi
environment.yml conda
  • conda-lock
  • ipykernel
  • jupyterlab
  • numpy
  • pandas >=2.2
  • pdbpp
  • pip >=21.1
  • plotly >=5.2.0,<6
  • pre-commit
  • pyreadr
  • pytask >=0.5.0
  • pytask-latex >=0.4.2
  • pytask-parallel >=0.5.0
  • pytask-r >=0.4.1
  • pytest
  • pytest-cov
  • pytest-xdist
  • python 3.12
  • r-forcats
  • r-plyr
  • r-precommit
  • r-yaml
  • statsmodels