Part A statistics project
As you know, this week we will be completing Part A of the Course Project.
This portion of the project involves using descriptive statistics and basic graphs to analyze
some data from a sales call center. This assignment uses only ideas from Week 1, so
you can get started on it right away.
First, a general comment about writing:
As with nearly all papers written for Keller courses, the goal here is to write with a professional
tone, appropriate for business. This really should look like a business document, prepared
for the client company.
Obviously this includes writing in complete sentences with proper punctuation and spelling.
But in business writing, there is also a real premium placed on being concise. That is, longer
is usually not better. Your clients and supervisor do not want to sift through pages and
pages of unnecessary charts and graphs.
In the past, I have had “A” papers that were 5 pages long, and “C-“ papers that were 20 pages long.
So, be sure and include only graphs and charts that are actually used – and proofread your
paper before submitting. When revising your document, give some thought to the overall
organization as well (nobody should submit a first draft).
Moreover, you do not want your paper to read like a “checklist”; that is, do not label sections
with things like “First Variable”, “Second variable”, etc.
Remember, this is a business document for a client, not just a homework assignment.
Next, let’s be clear about the required format:
Format for report:
- Brief Introduction
- Discuss your 1st individual variable, using graphical, numerical summary and interpretation
- Discuss your 2nd individual variable, using graphical, numerical summary and interpretation
- Discuss your 3rd individual variable, using graphical, numerical summary and interpretation
- Discuss your 1st pairing of variables, using graphical, numerical summary and interpretation
- Discuss your 2nd pairing of variables, using graphical, numerical summary and interpretation
- Discuss your 3rd pairing of variables, using graphical, numerical summary and interpretation
As stated here, you are going to analyze three individual variables and three pairings of variables.
I.e., you are not going to analyze all five. The individual variables you look at are completely up to you.
=> For an individual categorical variable, you will discuss percentage breakdowns; and include a
pie chart and/or a Pareto chart to support your description.
=> For an individual quantitative variable, you should describe the center and dispersion.
Decide which of the measures of center is most appropriate – either mean or median:
— If you use the mean, then you would describe the dispersion using the standard deviation.
Here, either a stem/leaf display or a histogram should be displayed to aid your description.
If appropriate (i.e. if the data is reasonably symmetric and bell-shaped), then you would also
the Empirical Rule to describe the dispersion. E.g., include a statement like
“ Based on this sample data, we estimate that about 95% of the data in the population lies
between __ and __ ” .
— If you use the median, then you would describe the dispersion in terms of quartiles; i.e. the
five number summary should be shown, and a boxplot used to aid your description.
Again, a couple of sentences explaining the five number summary would be given.
NOTE: Each of these analyses should require only about ½ to 1 page.
For the pairing of variables, the analysis will depend on the type of variables being compared.
Since only one of our variables is categorical (qualitative), we cannot do a pairing using two
categorical variables. This leaves two possibilities:
If pairing two quantitative variables, then you would obviously want to compare the centers
(e.g. compare the means or compare the medians) and also compare the amount of dispersion
in each set. But for two quantitative variables, it is also useful to include a scatterplot; this
will allow you to describe the correlation (strong or weak, positive or negative) between the variables.
If pairing a quantitative variable and a qualitative variable, you could compare the means between
categories. For example, you could compare the mean credit balance among rural, suburban and
urban residents. You can show back-to-back stem/leaf displays or side by side boxplots to aid your
Again, each of these analyses should require only about ½ to 1 page.
Including the introduction and conclusion sections, the entire paper should be between 5 and
10 pages in length.
Finally, a word about academic integrity. Students should submit only their own work, and
should acknowledge all sources. On this project, it should not be necessary to consult any
outside sources (you do not have to cite the text or Excel when doing basic calculations).
Copying from outside sources (e.g. internet sources or from other students) is not allowed,
and is a direct violation of Keller’s academic integrity policy.