Benson // w1d4

Winter 2015
01/15/2015

Planned schedule and activities

9:00 am: Morning haze, sleepwalking

9:15 am: Meet Paul, one of your TAs

Today will mostly challenge / exploratory analysis.

9:45 am: Hypothesis testing lecture

10:40 am: Short break

10:50 am: Hypothesis testing with MTA data

12:00 am: Food is the food of the soul

1:30 pm: Back to work on Benson

This afternoon, you are working on Benson. Tomorrow after lunch, we will start with presentations, so by then you should have your proposal and presentation ready. This means you want to start writing your proposal today.

5:00 - 6:00 pm: Vanishing point

Lecture Notes

The following ipython notebook is the completed notebook we did in class today, with all 3 parts. The pickle file that we used in parts 2 and 3 is on our private github repo.

Benson Distributions Hypothesis Testing with Bootstrapping 150115 .ipynb (102.9 KB)

Structure of the Benson Proposal

Proposals do not have very rigid structures and many differ from each other. Keep in short: ~ 2 pages is a good rule of thumb. Most clients won't be super excited by the prospect of reading a 19 page report. Avoid dry, passive voices. Don't put any code in. You want to excite them. Keep in mind that they are not data scientists, so keep the level high and not too technical.

Here are very general guidelines:

You can start with a very brief (2-3 sentences) problem statement, defining the question you're attacking. Then you can describe what immediate insights you pulled out of an initial, simple exploration of the mta data. You can put in a few graphs.

Then you can go into your plans of getting external data and analysis to help them solve their problem. You can put a sketch or two in there, or not, up to you. And you're done!