1. Identify spikes (outliers) in the data where extreme (high or low) sales values occur and correlate these spikes with actual calendar dates in 2003 or 2004 and with any holidays or special events or abnormally slow periods that may occur during these periods.

2. Modeling the data: a. Generate linear, quadratic, cubic, logarithmic, and exponential models. Output at most two models on any graph.

b. When generating the least squares models for this data, output the model and the R2 value and discuss these results.

c. What are the marginal sales (derivative, i.e. rate of change) for this department using each model. Discuss with detail what the marginal sales for each model indicates.

d. Compare your models. Which do you feel is best?

e. Remove appropriate outliers as you deem necessary for your favorite models and rerun the appropriate least squares model. What is the marginal sales and discuss improvements.

3. Comparing models a. Based on all models run, which model do you feel best predicts future trends? Explain your rationale.

b. Based on the model selected, what type of seasonal adjustments, if any, would be required to meet customer needs?

4. For the model selected as your preferred predictor, compute the percent rate of increase y2-y1/y1 for the next four weeks and provide appropriate backup computation. Note that week 91 sales is the last sales data available, so use your model to predict sales for week 92 and then compute the percent rate of increase. Repeat this process for y3-y2/y2 , etc.