Notebooks, Data, Code

Shortcuts will be availalbe here to all available datasets and computer code.

Notebooks

1. Equations on how to calculate milligrams of morphine equivalents
2. Dix Hospital Ledger data cleaning and basic descriptives

Data

These datasets are open access and available to the public. They contain no personally identifying information. All datasets are either used with permission or only contain public data. Attributions to orignal source material are provided. We respect data use agreements. Datasets are either used with permission or only contain publicly available records. They are available under various open access licenses and contain no personally identifying information.

1. Corpus of FDA-approved controlled substance US drug labels
   Published: October 3, 2019 by Nabarun Dasgupta
   Observations: 2,177
   Format: ZIP archive of .txt files
   Size: 50 MB
   Origin: United States Food and Drug Administration via NIH National Library of Medicine DailyMed API service License: Creative Commons Sharealike 4.0 International
   

2. Inactive Ingredients in Controlled Substances
   Published: November 14, 2019 by Nabarun Dasgupta
   Observations: 18,876
   Format: .xls codebook
   Size: 8.8 MB
   Origin: United States Food and Drug Administration via NIH National Library of Medicine DailyMed API service
   License: Creative Commons Sharealike 4.0 International
   

Code

Each set of code listed below is provided for public use, with the author's consent. Each code set contains a license that details sharing and reuse permissions. Code is written in SAS, Python, Stata, and/or R.

1. PubMed Search Volume Query
   Python 3.7, November 2019, MIT License
   By Nabarun Dasgupta
   This straightforward code will let you take a list of potential search terms and return the number of search term results from PubMed. It was create to identify noisy and irrelevant terms when building a search strategy for a systematic review.

2. Distinguishing Brand Versus Generic Controlled Substances
   Stata MP 16, October 2019, MIT License
   By Nabarun Dasgupta
   This code differentiates brand names from generic names of controlled substances. Useful for cleaning data from prescription monitoring program (PMP or PDMP), insurance claims, electronic health records, etc.

3. Data Cleaning for Historical Dix Asylum Intake Ledger
   Stata MP 16 in iPython notebook, November 13, 2019, MIT License
   By Nabarun Dasgupta
   This code will process, format, and create variables from the Dix Asylum (Raleigh, NC, USA) Intake Ledger, covering years 1856 to 1919.