Posted on August 12, 2014
The Wild Blue Yonder
So today is my second day off after quitting my job!  I’m very excited about it… my most recent client was very interesting, and I think we accomplished some really cool things while I was there, but getting it to the next step was out of my hands and it was time for me to move on.
Now, of course, I have to figure out what to do next!  Here’s the short list… I’m publishing this mostly to set a benchmark expectation and see how I do in the coming months, so it’s more a hopeful set of guidelines than specific goals.  Also, I may throw much of this aside if the perfect job shows up, but since I’m not actively looking that may take a while.  Read More
Posted on March 28, 2014
Fridays Off! And a movie Review…
I’ve gone to working part time, as of last week. After leaving Deloitte and the Air Force, and the fantastic team I helped build and was a part of up there, I have some higher standards for the work I do, and the contract work I started to help get us moved back to Louisville just isn’t as fulfilling. Don’t get me wrong, I’ve stayed a year longer than I expected to, and there are some great people and good potential… plus it’s been nice getting health care experience, but there’s nowhere to go in this position without having to step on some toes, and I have more I want to accomplish outside of work.
So here I am, at home, on a Friday. I’ve already visited Ollie’s Trolley for lunch (they’re only open weekdays during work hours), so that’s great!
Last week, I was mid-work on a bathroom renovation which, curiously, is still going on. But I’m almost done, really, I just need a tall person to help me with the top row of tile (*ahem*Angela*ahem*). I do have a ladder, just in case.
- March 2014 Bathroom Tiling
Posted on February 2, 2014
The 33 — reviewing a new series by J.C. Hutchins
Addison Creel made a deal with the devil (whether literally or figuratively we’re not quite sure — there is a character named Azael), and the bill is now coming due. 33 weeks of service to a secretive organization that has the ability to seemingly transport someone across the world in the blink of an eye, and who don’t seem to take “no” for an answer. Such is the plot for “The 33”, an episodic series coming from J.C. Hutchins, who wrote the 7th Son trilogy
Episode 1 is an introduction to a handful of characters and a teaser surrounding the modern-day sci-fi setting that is reminiscent of Neal Stephenson and other modern sci-punk-meets-real-world crossovers, and it certainly leaves one wanting more. The extent of the tech is also hinted at, but it’s clear Hutchins’ has more up his sleeve. This universe with nanotech pumping up people’s abilities, teleportation of some sort, holographic displays, and people who apparently have secret treaties with mars, is going to take some exploring.
Can’t wait for Episode II!
Posted on November 21, 2013
Efficient Geographic Coverage in SQL Server
Part 0: The Problem…
A fairly common problem in the real world is to make sure you have enough of something — hospitals, gas stations, water towers — to accommodate the need in a particular area. In Health care, where I finally got to tackle this problem in the real world, this is called “Network Adequacy”. In most places, the metric is basically the percentage of insured people that are within a certain distance of medical providers. The requirements change based on the provider type too… it may be important to have 90% of the population within 60 miles of a hospital, but you’d want an urgent care or primary care facility within, say, 20 miles of that 90%. The problem is compounded when you have to take into account the specialty of the providers. Being within 20 miles of a dentist, but 100 miles of a pediatric doctor or an OB/GYN is not sufficient.
It’s important to note that the distance is typically computed in actual travel time; it doesn’t help to be directly across the river from a hospital if you don’t have a bridge or a boat available. That version of the problem is wildly complicated but can be solved by tapping a travel time API such as google maps. These are complicated calculations; determining the distance between two points (a member and a doctor) in one second is probably doing pretty well.
But we have to do this for hundreds of thousands of members and tens of thousands of providers, for each provider type and specialty. That’s millions upon millions of seconds… we don’t have that sort of time. So we have to find a way to quickly pare down the results. The first and most obvious way is to eliminate the “drive time” requirement and focus first on the as-the-bird-flies distance… making some assumptions based on fastest interstate speeds, we know at least that we can’t do BETTER than a straight line, so at least we can limit the number of driving paths we have to calculate, and this is something we can do in pure SQL.
Read MorePosted on July 29, 2013
Adding a quick lunar calendar to the data mart
I ran across an article at Popular Science citing research that correlated Heart Surgery outcomes to the Lunar Calendar. That’s right… the claim is that you are more likely to survive and to spend less time in the hospital recovering from heart surgery if you get it during a waning full moon.  We know the moon can affect us in ways that make sense… for example an article popped up this week about the full moon disturbing sleep — it’s brighter, so it’s harder to sleep.  Ok.  And crime increases during full moons could make some sense too due to increased night visibility or such.  But medical outcomes?
Well, this gives us something new to play with, so let’s find out! Â First, I’m working on the data warehouse I’ve been working on for the last few months (wow, almost a year). Â It has a well behaved date dimension table, so we’re going to add a phase-of-the-moon calculation. Â The basic logic for doing that is:
- Lunar month is 29.530588853 days
- Jan 1, 1900 at 13:52 PM GMT was a full moon
- There are 9 indicators of the Lunar cycle, but it basically has 4 phases:
- Waxing Crescent when Lunar month is in the first quarter (29.530588853/4)
- Waxing Gibbous when Lunar month is in the second quarter
- Waning Gibbous when Lunar month is in the third quarter
- Waning Crescent when Lunar month is in the last quarter ( > 29.530588853*3/4)
- The other five are specific indicators, which we will round to the nearest day:
- New Moon is day 1
- Dark Moon is day > 28.53
- First Quarter is day between 6.88 and 7.88 (29.530588853 / 4 +- 0.5)
- Full Moon is day between 14.265 and 15.265 (29.530588853/2 +- 0.5)
- Last Quarter is day between 21.648 and 22.648 (29.530588853 *3 / 4 +- 0.5)