The second day began with a nice 10:30am start, but the late wake-up time still didn’t prepare me enough for that day’s statistics lesson. Today, Dr. Laber introduced us to the programming language R, and gave us several examples of how it can be used. We first looked at a massive spreadsheet of data taken from about 20 different dogs suffering from induced paralysis (from a getting hit by a car, falling, etc.), with cells including the time since incident, presentation, gender, breed, location of source, and biomarker proteins such as pNHFP and S100B that were measured. Dr. Laber asked Suki and I to find the primary causal factor that would determine a dog’s ability to walk after both 6 weeks and 6 months before setting a timer to 5 minutes and stepping out of the room. While a little daunting and stressful, we were able to deduce that if the average level of protein GFAPP exceeded a level of 0.31, the dog would have roughly a 92% chance of remaining paralyzed. He agreed, then showed us how R could be used to come across the same conclusion by comparing statistical algorithms of recovered vs. non-recovered dogs. He continued with these short exercises where he’d give us time to converse before explaining it in code, with examples such as latent states of depression of terrorists vs. non-terrorists, as well as methods of finding demographics of a population with just a small sample. It was quite the hour-and-a-half long hyper-speed presentation of applications, but our work felt important and our conclusions felt meaningful. Later that day, we went to the Laber Labs at BOM to meet with the team’s graphic designer, Lisa, to speak with her about a potential project for Friday using 3D imaging before we helped train the chess-bot Nona through about 600 more actions. Our time today felt pretty long, but the lab is full of so many interesting people and projects that each day is something completely new.
