Indiana’s Fight Against Opioid Abuse – Big Data

September 14, 2017

This is my take on Wired’s article posted earlier today.

Usually I can expect bad news when hearing about the opioid crisis in my home state. However, I was very pleasantly surprised to find out that we’re leading the charge with not conservative legislation but data analytics.

Utilizing SAP (I’m assuming) HANA’s ultra-fast in-memory database, the state is combining data from police (arrests) and hospitals (Overdoses, addictions, etc.) in addition to government held data. Three sources doesn’t seem like much but imagine sifting through millions of lines in Excel searching for clues – you wouldn’t get anywhere as Excel can barely handle 100k lines.

The in-memory database is the equivalent of being able to use notes on a memorization test. The database knows where to look, where to pull from, and how to write answers without being obvious to the teacher by looking elsewhere. Using non optimized hardware will make queries take exponentially longer in addition to costing more compute time creating unneeded overhead.

By connecting this data on an in-memory database the state can calculate queries that link objects together in mere seconds. This is where a data engineer or scientist can step in to make correlations between the data.

If person A has been arrested 2+ times in (County X) they are (X) times more likely to be an abuser.

Take the above example as being a correlation. Once these have been defined, the team can now start linking multiple correlations together. This can be done manually through self-observations but as of lately, you can automate AI or scripts that will find correlations for you. Maybe abusers have higher resting heart rates at a certain period. There’s millions of variables to test but by starting the process of linking data the state is already ahead of the curve.

Tableau’s interface.

Once found, the data can be dumped into a visualization tool which in this case is Tableau. Now you can tangibly visualize trends based on the links your engineer formed earlier.

Add in predictive based learning and you can start seeing how this is a very effective and powerful tool for many areas. In my experience, it’s connecting many business data sources together to optimize supply chains saving millions of dollars for the bottom line. However, in this case the opportunity is much greater – saving lives.