Why I Data, Part 1: Insurance

Welcome to my blog about all things data. I work with data every day. Generally, I do this work and then close my laptop and get on with whatever else is going on in my life. Sometimes, the work ends up being somewhat of interesting, and I wanted a place to write about it. I had used twitter to an extent, but I desired something more long-form. That’s what this blog is for.

The first 4 posts of this blog will be a series describing how I landed in the world of Data Science; from the beginning, as an intern at an insurance company doing data entry, to machine learning at BookLamp, A/B Testing at Apple, and ultimately to the present chapter in my life of data.

Alright, here goes…

In 2003, I was a Freshman scholarship forward for the St Bonaventure University Men’s Soccer Team, and I tore my ACL playing a spring game against the Rochester Rhinos (an A-League, professional soccer team in NY). I had dreams of a professional soccer career, and – with enough hard work – I believed I had the talent, determination, and genetics to make it happen. Dreams…

As a sophomore, I played the 2003 NCAA season with my still torn ACL. I scored 2 goals and had 2 assists on the season, and made the hardest decision of my life. After 16 years on a soccer field, I decided to call it quits and go home to Tallahassee, FL to attend Florida State University.

I enrolled at FSU and decided to study Economics. During my time there, I started working at a company in called Hunt Insurance Group of Tallahassee (HIG).

I began work as a bookkeeper in the accounting department. My mother was an accountant there, and she got me the interview to be a $7/hr Intern. I did data entry… all… day… every… day. It was the perfect, mindless job for the state my life was in at the time.

I was the human conduit for data that represented 100s of millions of dollars in company investments, which is kinda neat. Data entry turned out to be exceedingly boring. I was constantly thinking of ways to make this work more efficient. Still boring though.

In 2007, I graduated from Florida State University with a BS in Applied Economics, and was accepted into the MS Applied Economics program at FSU.

At this point, I had moved upstairs at HIG to the claims department. This was more interesting, and I remember the first project I worked on that resembled what we would now call “Data Science“.

HIG managed the Florida Sheriff’s Self Insurance Fund (FSSIF, renamed in 2011 to Florida Sheriff’s Risk Management Fund), and in 2008 my superiors came to me and said, “Sheriff [somebody] has asked if we could offer lower premiums and deductibles. Here is the loss data – see what you can do.” I looked at this data and thought, “Whoa!” Tens of millions of dollars in losses… boats, cars, trucks, ATVs, among other things; going back many, many, many years. “This is a serious spreadsheet”, I thought.

A wide-eyed grad student, I felt I was capable of practically anything, so I dove right in and Excel‘d the heck out of that data. I recall generating multiple linked tables that ultimately populated a final table of (Premium,  Deductible) combinations that would cover similar expected loss outcomes to what had happened in the past. The idea was to allow the Sheriff choose a (Premium, Deductible) option that best fit his preference, all of which would have effectively the same expected loss outcome.

I presented this loss data analyses to my superiors and was then asked to present it to the corporate office. HIG had since been acquired by Willis Group, so I presented to Willis corporate in a teleconference call. I recall the corporate executive stating over the phone, “This is great work. You are not going anywhere, are you?” I responded with a knee-jerk, “No, Ma’am.” That’s the first I realized that I had to tell someone that I was planning to move to Boise, ID after graduation to work for a small technology startup called BookLamp.

I informed my superiors about BookLamp and they were not all that happy, but they were still supportive. My parents had a similar reaction, and were also (ultimately) supportive. I graduated grad school and was ready to test myself in Boise, ID – of all places. I remember thinking, “I never would have figured I’d somehow end up in Boise f#%^ing Idaho. How odd.”

This was just as Hal VarianGoogle’s Chief Economistremarked, “the sexy job in the next ten years will be statisticians.” Loud and clear, Hal. I could hear you loud and clear.