Part D.

Suppose you have sales both near and not near to the plant as well as sales before and after the spill. Suggest an alternative strategy to estimate the effect of the oil spill on housing prices that is preferable to both MA and MB?

Word Limit: 800 words for entire question (i.e., all subparts).

QUESTION 2

Imagine you work for a large department store, which highly values customer service. The following chart shows how customers contact the customer service centres.

You begin to discuss the chart with your manager. Immediately, she has the following queries: “I want to see the overall trends, but it is difficult to see with all the seasonal spikes in the time

series. I’d like a simpler view into the trend.” You decide to create some charts to address your manager’s queries.

Part A.

Using the four frameworks typology, identify the type of chart you would use to address the query and explain why.

Part B.

Sketch two alternative charts for the query. For each chart, provide a brief explanation of your design choices. To sketch the chart, you can use any tool you want (e.g., you can use a          software tool likeinfogram, excel, or R). Alternatively, you can sketch the chart using pencils,  pens or markers on paper, then take a picture of the charts and paste them into your solutions document. You can access the underlying data “customer_service.xlsx” on Ed.

Part C.

Evaluate your two charts and explain which you would select to further develop to present to your manager.

Word Limit: 500 words for entire question (i.e., all subparts).

QUESTION 3

As part of a preventive health program, we are interested in building a model for predicting diabetes among women. We have access to the following information for each woman:

-       diabetes: Yes or No for diabetic .

-       npreg: number of pregnancies .

-       glu: plasma glucose concentration

-       bmi: body mass index

-       ped: diabetes pedigree function

-      age: age in years.

We have split the available data into a training dataset with 332 cases and a test dataset with 200 cases.

With reference to the above case, please answer the following questions:

Part A.

We have run a logistic regression to predict diabetes using npreg, glu, bmi, ped, and age. The R output from this logistic regression is as follows.

Based on the output, write down the mathematical equation of the logistic regression associated with this R output.

(2 marks)

Part B.

Based on the output in Q 3 Part A, provide an interpretation of the coefficients associated with bmi and npreg?

Part C.

Based on the output in Q 3 Part A, explain and justify the predictors that have a statistically significant relationship to the response?

QUESTION 4

As an alternative to the logistic regression in Q3, we have also fitted the classification tree below:

Part A.

Based on this decision tree, would you predict that a woman with the following characteristics has diabetes: npreg=0, glu=140, bmi=47.9, ped=0.259, age=26? Justify your answer.

Part B.

Provide TWO facts to demonstrate the consistency of results from the classification tree above with the findings reported in Q3 Part A.

QUESTION 5

You are a data analyst for AppCo. AppCo produces a smartphone app that allows users to      virtually try on clothes. It is funded by having sponsored links to online clothing retailers.          AppCo tells its users, “top brands, all sizes, best prices” . AppCo has a dashboard that is used by Board members that shows sales, revenues and a comparison of sales by retailer. You       have been asked to consider whether having a selection of the most popular sizes at a slightly lower price would increase revenue. The CEO has said, “well just use some of your magic A/B testing …” .