Fall Semester

I blame my lack of blogging in recent months on being busy with my first semester of graduate school. In the interim before I start the next semester, I want to let you—oh loyal reader—know what was going on in my life during that time.

1. Per my stake president’s counsel given more than a year ago and which I sadly only barely got around to heeding, this semester I went on a date a week. This was a fabulous experience. Going on more frequent dates made me less nervous about going on any particular date. It gave me more opportunities to get to know some really amazing girls. I’m often amazed at how such excellent girls are willing to give me some of their time. It’s been a pleasure learning about them and getting better at letting them learn about me.

2. I worked really, really hard doing data mining for the BYU bookstore as part of a class project. I got to be really passionate about this, as boring as it might sound on the surface. The bookstore gave my group a large amount of sales data from their website over nearly the past decade. Our task was to turn that dry data into actionable knowledge about the bookstore’s customers. That challenge is somewhat akin to an archaeologist being trying to reconstruct the daily lives, beliefs, and values of an entire civilization given a site full of pot shards and millennia-old garbage. The only way to make the bookstore data really useful, it seemed to me, was to leverage every possible datasource in existence. To that end, I augmented the bookstore data with the following datasources:

• Yahoo!’s Geoplanet web service: Used to geolocate the billing and shipping addresses.
• Yahoo!’s Geoplanet Data: Provided a hierarchy of placenames and unique geographic identifiers.
• United States Census Small Area Income and Poverty Estimates: Provided median income data for all United States counties.
• ANSI FIPS Codes: After geocoding these state and county identifiers, this dataset allowed us to reference the US Census data which uses FIPS identifiers natively.
• tz_world Timezone Shapefile and Index: Used to determine the timezone each address was located in.
• Google Trends: Provided search frequency data for searches relevant to the bookstore’s business from January 2004 to November 2009.

I also worked on using weather data to determine what the weather was like at the place and time an order originated from, but this was too time-consuming so I had to drop it. Yet I still think it would be interesting to see what correlations you would find between weather conditions and people’s desire to shop online. If the NOAA would make their historical observation and model data available via their web services this would be trivial.

Anyway, our final report is here. It’s not the best-written thing, but the pictures are at least interesting!

3. I kept up a fairly full schedule with my dinner group, poetry club, Institute classes, some running, tennis, one game of racquetball, hiking, camping, helping at a school, playing trombone in a pit orchestra, occasional family history research (that has certainly suffered since returning to school), reading about all kinds of random things in the library, and generally being really bad at replying to phonecalls.

Overall it’s been a really happy time. School’s stress has usually been manageable, the projects and homework often enjoyable, and the material just plain cool. Dating, for perhaps the first time in my life, became more enjoyable than overwhelming. Things unmanageable became manageable because there was usually somebody to talk to when life was perplexing me. My bishop has provided sound counsel and inspired blessings, helping me feel more connected to God by means of one of his servants. And, as if things couldn’t get any better, the price of cheese went down, small children kept on bravely facing the world, and the sun returned from its absence every day. Boy, life is good!