Time Series Analysis Outpatient Pharmacy RX Volume

The following was submitted as a final research project for a capstone course offered through the M.S. Data Science program at the University of Wisconsin - Eau Claire.

Data on daily prescription volumes (totals) was obtained for three leading outpatient pharmacies in the NWWI and analyzed with descriptive and predictive methods. Descriptive methods displayed totals over time in multiple figures along with wilcoxon signed rank paired tests to illustrate patterns and trends since 2009. A neural net, Facebook's "Prophet", and a TBATS model were then fit and assessed to obtain a model offering accurate predictions on totals 2 weeks (13 days) ahead of time.

Data Cleaning

Data cleansing was carried out prior analysis with cleaning_phase1.py and "new data only.py" files.

Project Contents

1. Data loading and exploratory analysis
2. Descriptive analytics and statistical testing
3. Predictive analytics
4. Final predictions

See TSA_Outpatient_Pharmacy.md for markdown file containing all project components.