Mathematics for Scientists and Engineers

Math 530 Fall 2011 - Homework -

Best Approximation Problem Using NOAA Climate Data - Due Wednesday Sept. 14

(i) Find the best linear and (ii) quadratic approximation to a real dataset, plot your results along with brief details of your work.

You may use any dataset you'd like, as long as you can identify an independent and dependent variable. Below, you can find details on how to obtain the climate data we discussed in class.

Get your hands dirty with some real climate change data.

This readme file will be helpful. You can download the maximum monthly temps dataset directly here, (this is the dataset I used in my example in class). Or, if you want to work with a different, but related dataset of average monthly temperatures or precipitation values look here after peeking at the readme. The main website for all of this data is here.

Helpful Hints

Import the text file into MATLAB using the "load" command (>> help load). The first column of data contains both the year and the location associated with each row. (See the readme). After 117 rows the location (sensor) changes so if you want to keep it simple, you can just worry about one location.

This means, if do:

>> A = load('mydataset.txt'); --- then, MyData = A(1:117,:);

Columns 2-13 refer to January through December, and column 14 is the mean. Make a seperate approximation for each column. You might need to "clean" data that has obvious errors, but indicate how you do this. The first few pages of Gockenbach's MATLAB Tutorial will be helpful if you're having trouble getting started. Start this weekend, and feel free to ask me questions.