PeerTax

This is a Data Science project looking into the Peer Review Taxonomy.

As open data sources it is looking at the reviews from F1000 and validating against the BMJ Open.

Setup

The project provides configuration for either using a Python virtual environment or Docker.

Setup using Virtual Environment

Pre-requisites

• Python 3

Create Virtual Environment

make dev-venv

Run Tests

make dev-test

Start Jupyter

make dev-jupyter-start

Setup using Docker

Pre-requisites (Docker)

• Docker and Docker Compose

Build and start Jupyter

make jupyter-start

This will build the Jupyter image and and start it via Docker.

Run Tests (Docker)

make test

Project Structure

The project is structured in the following directories:

• data: data downloaded from other sources
• LDA: Notebooks related to LDA
• notebooks: Notebooks related to preprocessing the text (sentence and token splitting)
• peertax: Python package used by the notebooks
• pickles: Intermediate output files
• scripts: Other scripts (to run LDA models in parallel)
• tests: Python tests for the peertax package