/pre-doc-record-linkage

Primary LanguageJupyter Notebook

Record Linkage

In this repository

  • This readme file contains the agenda for the 3-day training session
  • Notebooks - example notebooks we use in the training
  • Presentations - presentations (will be added)

set-up

The example notebooks use the Python recordlinkage package. You may be able to simply pip install recordlinkage to add it to your local machine

Optionally, anaconda provides a nice package management system and you can create virtual environments for different types of analyses. This conda cheatsheet includes commands for setting up and creating virtual environments.

  1. Create named env: conda create --name py3-rl python=3.6
  2. Activate new env: source activate py3-rl (mac) or conda activate py3-rl (linux/windows)
  3. Add recordlinkage: pip install recordlinkage
  4. Add ipykernel for step 5:conda install ipykernel
  5. Make kernel of new nev available to jupyter python -m ipykernel install --user --name py3-rl --display-name "py3-RecLink"
  6. Launch Jupyter notebooks: jupyter notebook

Once you've run the above once, you should now have a py3-RecLink kernel available in your installation of Jupyter notebooks. If you need to add additional packages, you may need to repeat steps 2 (to reactivate the kernal) and steps 3 or 4 (to install the specific package(s)).

NOTE for the ML notebook example the sample data is too large to store on github, so we stored it on a Drive folder. You can download it here

Additional resources

  • Patentview-API repository teaches API calls to pull PatentView data, and is built to work using binder so you do not need to install the required Python packages (pandas and requests)