Exam Scores Response Time ms Network Traffic

55 5 12
60 6 13
62 7 13
65 7 14
67 8 14
68 8 15
70 9 15
72 9 15
73 10 16
75 10 16
76 11 16
78 12 17
80 13 17
82 15 18
83 18 18
85 22 19
87 28 20
88 35 25
90 50 40
92 75 65
75.4 17.9 19.9 MEAN
STADARD
10.55511849 17.5346274 12.21259057 DEVIATION
-0.194315242 2.320312551 3.140958074 SKEWNESS
-0.854624184 5.560376971 10.4720593 KURTOSIS
Lab Analysis:
Skewness and Kurtosis Interpretations

1. Skewness and Symmetry After calculating the metrics for each set, Dataset 1 (Exam
Scores) is clearly the most symmetric. Its skewness value of -0.19 is the closest to zero. This
suggests the scores are distributed pretty evenly on both sides of the mean.

2. Positive Skewness

Dataset 3 (Network Traffic) is the most positively skewed with a value of 3.14If you check my
histogram for this data, you can see exactly why: most of the data points are bunched up on the
left side , while a long tail stretches out to the right because of those high spikes like 40 and 65.

3. Kurtosis Values For this part of the lab, Dataset 3 ended up with the highest kurtosis value at
10.47

4. Peaked vs. Flatter Distributions

Dataset 3 is the most peaked since it has that very high positive kurtosis. On the other hand,
Dataset 1 is the flattest of the bunch (platykurtic) because its kurtosis value is actually negative
at -0.85

5. Comparing Shapes and Signs The histogram shapes and the skewness signs definitely match
up. Dataset 1 looks balanced, whereas the histograms for Datasets 2 and 3 both show that clear
right-hand tailing off that you'd expect to see with positive skewness numbers.
6. Server Response Time In a software environment, we expect server response times to have
positive skewness. Ideally, most requests are handled almost instantly (the peak on the left), but
you'll always have those rare "edge cases"—like network congestion or database locks—that
take much longer and create that long tail on the right.

7. Network Traffic Spikes High kurtosis is a major risk for network engineers. It basically
means the network is prone to sudden, extreme traffic "bursts" or spikes. If the infrastructure isn't
scaled to handle these outliers, the whole system could crash or experience massive latency, even
if the average traffic looks perfectly normal.

8. The Effect of Outliers Adding the extreme values (10, 150, and 200) to Dataset 1 changed the
results completely. The skewness jumped from near-zero to $1.88$, and the kurtosis shot up to
$6.60$. It’s a great example of how just a few outliers can totally warp your distribution and
make your "average" data look much more peaked and skewed than it actually is.