FREQUENT OPIATE PRESCRIBER

The aim of this project is to reduce the death due to drug over dose. Here we are determining who is prescribing opiate very frequently in USA. We classify the prescriber frequent if he is prescribing more the 10 opiate associated drug within a year.

How to use it

1. I have used data from kaggle site This Link

2. Set up conda env:

   1. create an env

      conda create -n fop

   2. install all required packages

      conda env update -f env.yml

3. Notebooks:

   1. When I was implementing this project, I used notebooks. I have them in two files.
   2. In EDA.ipynb I have all my Data analysis, and in ml_model.ipynb I have written all the training codes.
   3. The config.py has all hardcoded variable values.
   4. The pipeline.py has the main pipeline
   5. The preprocessors.py has all the custom preprocessors.

4. package_fop folder:

   1. This folder has all the python scripts for deployment easy style.
   2. I have created the package of this package, so that later I can easily import it in Flask app.

5. setup.py file:

   1. To build the package use

      python setup.py sdist bdist_wheel

   2. To locally install type below command from the directory where setup.py is present.

      pip install -e .

6. MANIFEST.in

   1. This file contains the name of extensions which are not by default added to sdist/bdist. Because of the names the files get included in package.