Question

2. For the variable you identify, construct a graphical dis-play. What conclusions can you reach from this other plot that cannot be made from the boxplot?

The Variable I chose was wrong billing

3. Summarize your findings in a report that can be included with the task force’s study.

Answer 1

Frequency of types of CS errors 160 140 120 100 80 60 40 20 0

Costs of Type of errors 140 120 100 80 60 40 20 0 So, from the above two column charts it is quite evident that No two variables can be considered as normally distributed as the maximum frequency or the maximum cost does not belong to the median and mean, median and modes are different for the two variables. Also, from the shape of the column charts it can also be concluded that both of the graphs are approximately negatively skewed as both have left tails.

3)

The descriptive statistics table of "frequency of types of error" it is quite clear that the average of types of errors is 29.87 and the deviation from the mean of those types is 31.06. The minimum number of the errors is "subscription not ordered" type error and the maximum number of errors is "wrong billing error". The median frequency of errors corresponds to "termination error". The total frequency of all the errors is 448.

The box plot and the descriptive measures of "costs of types of error" it is clear that the mean of all types of costs of error is 46.75 with a standard deviation of 25.35 from the mean. The 1st and 3rd quartile corresponds to "wrong number of connections" and "wrong subscription type" respectively, while the median corresponds to "wrong end date". The cost is minimum for "incorrect accessory" type of error and maximum for "wrong billing".

No variables can be considered as approximately normally distributed as both of them are negatively skewed.