Me, My Soul, and I - Douglas Hofstadter has a new book, I am a Strange Loop. Must finish GEB asap.

How Pi Works - OK I am fudging this one a bit since I did not post anything on Pi day but I am going to edit the post date so it looks like I did.

Zelda Music of Golden Proportions - More math! Phi and the Zelda soundtrack. Via

Cool Numbers - Wednesday is Pi Day (3-14) and in honor expect a lot of math posts until then. Here is a list of one guy’s coolest 10 numbers. Fun nerdy read.

A couple of sites I just ran across that pwn your data visualization. Swivel and IBM’s Many Eyes. Check out the links but best described as data pr0n.

Where I work falls under state budgeting which means two things, everyone is underpaid and everyone’s salary is public information. While everyone is underpaid, some of us are more underpaid than others. So with budget planning beginning, I am going to my boss to justify a raise. Here is where is the second part of my first statement comes into play. It is fairly simple to setup statistical models to show how underpaid you are.
The model I am taking to my boss is a fairly simple one. Salary = Worth + Years Worked * Annual Raise. Here Worth is pretty much the intercept of the graph and represents how much each person is worth in relation to one another with Seniority factored out. A fairly simple model but somewhere to start and I also only included people in the pool with my job or similar. No point in comparing my salary to the directors’ salaries. Some reading about multiple regression: Intro to Regression and Multiple Regression.
Here is how I did this in Excel 2007. You’ll need the Data Analyzer plug-in installed. First you’ll need 3 columns of data - Pay, Years Worked and optionally Name. Go to Data Analysis -> Regression and select OK. Here you select your Y Range (Salary Column) and X Range (Years Worked). When selecting options, make sure you have Line Fit Plots also selected. This generates a histogram (wherever you specified your output range). The X-axis will be Years Worked with 2 columns for each data variable. One is Pay and the other is Predicted Pay. You’ll need to match up according to the correct Seniority values but you get the basic gist. If there are any math/stats wonks who would like to add input to this; I would love to hear it.

My favorite statistics blog, Statistical Modeling, Causal Inference, and Social Science, is running a slew of best of articles for this election day.

• Persistence of Red\Blue Illusions - Interesting slant on the perceptions of party leaders
• Civil Liberties and War - an analysis of a paper on that subject
• Ser-Venn-ity Prayer - Good old fashioned nerd humor
• Why it’s rational to vote - Excellent critique of the of voter apathy

Democrats gain 32 seats - Statistical predictions and analysis of the the upcoming mid-term elections.

One Million Ways to Die - Considering today’s date, an appropriate risk assesment of terrorism. “In fact, your appendix is more likely to kill you than al-Qaida is.”