WeatherPages
National Weather Service
Southeast Regional Climate Center
From each of these sites, you can download many different weather variables
for many cities in the U.S. I have the following data that I have dowloaded
from these sites:
Athens, GA 1994 - 1995
Chicago, IL 1994 - 1995
Athens GA 1990 - 1997
Chicago, IL 1958 - 1990
These are data files. To import these into Microsoft Excel, copy these files
from this page (or obtain your own data sets from the above web pages) and
follow these steps.
1. Open Excel and under the File menu Open the data file.
2. You will see a dialog box showing the suggested method of conversion. The line 'original data type delimited' should be selected. Click on NEXT.
3. Select 'comma as delimiter' and click on FINISH.
4. Excel will place the data into cells. You may need to resize the columns to see all the information in a cell. Also, near the top of the page is a key to the abbreviations that are used in the data set.
I have completed this procedure with the Athens, GA
1990 - 1997 data. This Excel file may be easier for you to use since
I have put the data into columns and I have found the monthly average of
the mean temperature. In another Excel file,
I have set-up the following monthly averages which I have graphed.
Since I know the temperature is a recurring cyclic event, I will model
it with the sine curve. So, I have graphed the equation y = a
sin (bx + c) + d and compared it to the graph of the actual data.
In my spreadsheet, I have used the formula for the temperature in January
as y = v23 * sin (v24 * w3 + v25) + v26, where a = v23, b = v24,
c = v25, d = v26 and the number of the month corresponding to the above
column (i.e., january was w3 = 1). I did this for each month where
a, b, c, and d were the same values but the number of the
month was different. I put these values into a column as seen below:
Then, I have set up the spreadsheet as follows:
Note, that because of the way that I set up the spreadsheet, as I change
a, b, c, or d, I can see immediately how this change affects
the function with respect to how it models the actual data. As you can see,
I can manipulate the data very easily, and keep changing the parameters
until I get a close agreement with the data. From my experience, the ease
of the set-up will appeal to students and will keep them actively searching
for a close agreement between the actual data and the equation modeling
the data.
Possible extensions to this would be to graph the average mean temperature
for one year and note that it is not as close to a sine curve as the average
over seven years. A discussion of why this is so could help students understand
data sampling and why statisticians seek larger number of data.