A data exploration, analysis and modeling project on employee and applicant data to explore, analyze and predict potential length of applicants' employment at given organizations.

Code Organization

• data_analysis
  • data
    • [file1].csv
    • [file2].csv
  • notebooks
    • data_analysis.ipynb
  • src
    • data_load_transform.py
    • visuals.py
    • model.py
  • .env
  • README.md
  • environment.yml

Setup a conda environment

To get started, setup a new environment (preferably with anaconda) using ./environment.yml file

conda env create -f path/to/environment.yml

Set the following environment variables with the right value in your environment.

The default values are as follows:

APPLICANTS_DATA=data_exercise_applicants.csv
HIRES_DATA=data_exercise_hires.csv
DATA_DIR=../data
REPORT_DIR=../reports

Data Munging and preprocessing

Data munging is done by running the data_load_transform.py module.

To run the data munging module, go into the ./src subdirectory and type the following in your terminal

python data_load_transform.py

This will,

• load the datasets (applicants and hires)
• preprocess the data (encoding target, binarizing categorical variables, filling nan values, etc)
• splits the data into test and train samples

Analysis

The analysis notebook can be found under ./notebooks directory as jupyter_data_analysis.ipynb Run the following command in your terminal from the ./notebooks directory

jupyter notebook jupyter_data_analysis.ipynb

Classifying Tenure Length

Here, we can train an algorithm of our choice and make predictions on our test data. To tune a model, train, test and return a classification report, run the following command in `./src' directory from your terminal

python model.py [[path to data directory | Default: ./data/]]