The objective of this notebook is to provide exploratory analysis for PT Utilization benchmarking. There are several goals of this project:
Provide a base of historical data on which the benchmarks will be derived.
This needs to be cut by different member demographic types in order to correctly use historical data.
A concept of seasonality would need to be introduced (Eg, Higher ED product utilization following the school calender)
Understanding how historical trends translate into "good" or "bad" utilization levels.
Product specific utilization metrics will need to treated seperately and will obviously need to be treated at a product level only.
Rolling up benchmarking metrics to the product level will be the most macro we anticpate being able to go. This is because metrics won't be applicable cross product.
That being said, we can roll up certain statistics, such as % of members in x percentile, to the vertical level.