CMS provides a 5% sample of their full claims data, but at prohibitively high expense both in time and cost. They have provided a synthetic dataset to allow for experimentation and rapid prototyping of solutions without the barriers associated with the offical dataset.
This repository provides the code to:
- Obtain and load the synthetic dataset into a mariadb database
- Perform descriptive analytics on the synthetic dataset
- Peform inferential analytics on the synthetic dataset
- Perform machine learning analytics on the syntheic dataset
The 20 samples can be obtained from the CMS site (see citations). These data samples can then be pre-processed to allow them to be read into the database. Each dataset can then be loaded into the tables in the database.
The files to setup the database on a linux server are present in the scripts_setup folder. These can be run by directing the .sql file to the mysql server. $ mysql < [setup_table_1_scpript.sql] There are also bash scripts to load the csv data. $ source ./[load_table_1_csv_data.bash]
There were minor naming issues, likely due to dataset updates, encountered when downloading the dataset and the resulting database counts should be compared against CMS Synthetic Data Description
[TODO]
[TODO]
[TODO]