Lecture 19
For today you should:
1) Work on your preliminary report.
2) Prepare for a quiz.
Today:
1) Survival analysis
2) Quiz
For next time you should:
1) Finish a draft of your preliminary report by 2pm Friday. If you have not already shared the document with Paul and me, do so immediately!
Preliminary report suggestions
1) Explain your project to a naive audience.
2) Between you and your partner, have a concrete goal for the preliminary report: what do you need from the liaison, and how does the report help you get it? Hint: what can you do to increase the chance that it gets read?
3) Don't complain about data that's hard to work with -- that's your competitive advantage!
You might like the following short talk, which hits many of the themes of this class.
And here's some optional reading about the impact of data science on insurance. While you are waiting to find your own unique authorial voice, you could do worse than to imitate the Economist:
A tricky business: Data-driven underwriting contains great promise and grave perils
Here's the intro:
INSURANCE rests on the idea of imperfect knowledge. Since its creation in the 17th century, insurers have sought to amass lots of policies in each class of risk they cover. They do so not only to make money, but also to be safer. This is the law of large numbers. Insurers don’t know exactly where risks lie. By insuring a barrelful of policies, it matters less if one of them is bad.
But the law of large numbers is threatened by the rule of precise data.
Thoughts? How about this figure?
