Econ 70 - Fall 2005: October 2005 Archives

« September 2005 | Main | November 2005 »

October 27, 2005

The Central Limit Theorem

The Central Limit Theorem can be viewed as the motivation for portfolio diversification.

One share each of two independent stocks:

Two shares of one stock:

A review of the definition of variance and covariance:

Another view of the Central Limit Theorem is as a way to approximate a binomial distribution for large samples.

Posted by bparke at 10:58 PM | Comments (0)

October 24, 2005

Stata 2

Due Tuesday, 11/8.

The Stata dataset website.dta records visits to www.econmodel.com over a
48 weeks period. Your assignment is to produce a report analyzing traffic at this web site.

To get started, click on the link above with your right mouse button. Choose "Save Target As". Save the file someplace where you can find it later. Start Stata 8. Choose the menu option "File | Open". Go find website.dta, and load it into Stata.

Once you have website.dta loaded, start "getting to know your data." Look at the raw data in the data editor. Generate some summary statistics. Make some graphs.

If you generate any transformed data, it is a very good idea to save your changed data in a new dataset using "File | Save As". If you make a mistake or if you just make a mess, you can then go back to the original data.

The variables are as follows:

day, month, year are the calendar date. dayofweek is 1 for Sunday through 7 for Saturday. week is 1 through 48.

visits is the number of different IP addresses that requested pages from the web site during a given day. pages is the number of web pages requested during a given day. hits is the number of files requested during a given day. The number of hits is typically much greater than the number of pages because each graphic element on a page is a different file. bandwidth is the number of megabytes requested during a given day.

For this assignment you will turn in a report with the following sections:

1. Introduction
2. Data Description (define your variables and transformations)
3. Empirical Results
4. Conclusions

You will focus on visits, but may also work with pages. You should check for the existence of the following patterns in the data and measure the magnitude of each effect.

1. Day of the week.
2. Growth over time using monthly figures.
3. Special days of the year.
4. Relation to the typical U.S. academic calendar.

What I am looking for is a few pages where you demonstrate that you can put together descriptive statistics and graphs. Think of yourself as a $100 per hour statistical consultant getting paid to analyze traffic at this web site so that the owner can plan for next year. At that pay grade, lots of charts and graphs are important to impress the client.

Posted by bparke at 08:21 PM | Comments (0)