An example solution to the stock price problem:
I used IBM monthly closing stock prices from Feb 04 back to Jan 97. After downloading from the finance.yahoo.com web page, these data occupied cells e2 to e87 of the spreadsheet, some of the values being 100 (Feb 04), 99.23 (Jan 04), 92.68 (Dec 03),...,143.75 (Feb 97), 156.88 (Jan 97). To compute the percent price differences per month, I highlighted the cells i2 to i86 and entered
=100*((i2:i86)-(i3:i87))/(i3:i87)
The percent changes then appear in cells i2 to i86. For example, the first two and final values were 0.775 (Feb 04), 7.067 (Jan 04), -8.369 (Feb 97).
A time plot, stem and leaf plot and histogram are shown in the next three plots. These were drawn using standard Excel options for the timeplot and histogram. For the stem and leaf plot, I chose to use just the whole number of percent (-46 to +35) treating the first digit as the stem and the second as the leaf. Note the use of two separate rows labelled -0 and +0 respectively for the values between -10 and 0 and those between 0 and 10. The drawing of the stem and leaf plot was facilitated by first arranging the data in order using the Rank/Percentile function.
-4 | 46
-3 |
-2 | 2
-1 | 000024589
-0 | 00001222223334444555677788899
+0 | 00001111222344444445567778
+1 | 0111114555677789
+2 |
+3 | 1 5
DISCUSSION. For this series, the main characteristic feature appears to be four outlying values (two positive,two negative) which stand out strongly from all the rest. Whether or not these are formally "outliers" in the sense defined in Chapter 5, they clearly stand out from the rest of the numbers. The time plot does not seem to show any longer-term trend or periods of high volatility; here as well, the main features that stand out are the four outlying values that correspond to sharp changes in the stock price.
NOTES ON THE GRADING. This question is worth 10 points; the other two questions on this assignment are worth 5 each. At least half the 10 points should be for accurate description of what was done, including details about which stock it was, what time period was covered, and how the percent change was calculated. The rest of the points are for the appearances of the plot and the discussion. I don't want to be fussy about precise details of how the plot looks, but it's important that it conveys the right overall impression. As for the discussion, it should highlight important features such as outlying values and periods of high volatility.