Friday, August 29, 2008




Well since I wanted to start a new thread, and you were virtually begging me to make another speadsheet/chart about some random nonsense here goes. I went to weather.com and got the average and record highs and lows for every day of the year. This data seems to range from about 1926 to about 2005 and is for West Berlin (probably for somewhere near here, but they don't tell you where).

http://www.weather.com/weather/climatology/daily/08091?climoMonth=1

In sundata you can see the sunrise and sunset times for the year, as well as the length of the day. Nothing to interesting here, just that it's not perfectly symmetrical.

Next in tempavg you can see all the averages/records for the year. Some interesting things here, first late January is the colder time of year, late July the hottest. Nothing shocking, the records though show something that surprised me. The difference between the coldest record low in January, and the warmest record low in July is about 75 degrees, compare that to the difference in the two extreme highs of only 40 degrees. In case you were wondering that huge cold spike is -25 on Feb 9th 1934. As you look at the lines for the records the two things that stick out are that the low has a sudden drop in mid January to mid February. Second is that the record high doesn't vary that much. This seems mainly because there seems to be something stopping it from rising much above 100. The best I could come up with for this would be because we have a large mass of cold area (Canada) near us, but we don't really have any large hot area.

Last we come to temprec, which list how many days in a given year hold the current record low or high. This chart doesn't have a lot of data and there is some chance that they moved the place where these records are recorded at some point, but I still think it shows some interesting things. First I notice that there are almost always more record lows than highs. Most of the spikes of highs have a larger spike of colds in the same year, or in the next year. The two times when there were more highs than lows is the first decade and last two decades of data. I think the first group can be explained because we know the high doesn't go that high ever, thus in the first few years they were recording the data most of the highs just haven't been beaten. However, I'd bet if you had 50 years more data into the past you'd see that group move to the start of that data. The recent data also shows some interesting trends. You can see there are very few lows, however, there are also either few highs or many highs. In other words it seems as though the last 20 or so years have been more mild, with occasional increased wild spikes of heat.

I'm not sure what any of this means, but I am certain it's ground breaking, and will be expecting the grant money any day now.

http://daleswanson.org/blog/temps.pdf

http://daleswanson.org/blog/temps.ods

Thursday, August 7, 2008

Voting

I was doing some thinking about voting. During presidential election years we generally have 50% - 60% turnouts. Some people don't like this, and wish it were closer to 100%. I don't think it's a problem though. A lot of studies have been done and most find that the voting population and the non voting population are similar, in other words even if 100% voted the winner would generally be the same. Really if you think about it, even 1% of the total population is considered great for statistical sampling, over half is almost guaranteed to be an accurate sample.

There are a few ways to increase turnout. Australia has mandatory voting, and has the highest turnouts in the world (>95%). However, I think freedom not to vote is important. Not only that but whenever you force someone to do something, you get a poor job (remember this for whenever you become a manager, forced weekend inventories = unreliable counts that need to be redone).

Another idea, which I found interesting is a million dollar lottery awarded to one voter at random. Arizona considered this about 2 years ago, and while I can find a lot of stuff about it dated from then, I can't find anything that would indicate if it actually passed or not.
http://www.cbsnews.com/stories/2006/05/23/politics/main1644607.shtml

While I was driving to Wendy's just now, I was thinking about this, and guessed that as intelligence or education increased likelihood to vote would increase too. Now this may seem fairly obvious, especially now that I know it's true. One thing I did think is that more intelligent people may become frustrated with the system. It was rather hard to find any good results. Almost everything I found had to do with comparing how smart people are to which party they vote for (most of which took the form obvious propaganda for whichever party the writer liked). However, I ended up finding a perfect answer in Wikipedia. It was from some book about India, and it actually shows some interesting things about India (voting peeks in the middle income and education), but for America it's quite clear that both high income and high education equate to being more likely to vote. It's probably worth noting that while there may be more higher educated people voting as a percentage of their group, there are a higher absolute number of less educated people in general, and voting.
http://en.wikipedia.org/wiki/Voter_turnout#Socio-economic_factors

Education Effect On 1988 USA Election
38% No High School
43% Some High School
57% High School Graduate
66% Some College
79% College Grad
84% Post-Graduate

I was surprised at how strong the correlation between higher education and voting was. However, it is always important to distinguish between correlation and causation. Higher education may cause likelihood to vote, but then again maybe the people who tend to be predisposed to pursue higher education may also be predisposed to vote. That being said I still believe that higher education increases likelihood to vote, and if not makes those who do vote better voters.

I think this is generally good. Less educated people are more likely not to vote, and they are more likely to vote for arbitrary reasons if they do. I've said this before (and you disagree with me), but probably the best return of an investment we can get as a country is to pay for education. A better educated populace is more likely to vote, and more likely to vote based on solid reasons if they do. It's easier to fool more ignorant people into voting for you, which is why we'll never see any real move to provide more education.