Statistics Homework

1

Written Assignment 1 (Group Assignment) Due uploaded in a single .doc, .docx, or .pdf file to Canvas by 11:45 PM on Tuesday January 26th You work for an automobile insurance company. Your boss has assigned you the task of reviewing recent insurance records for SUV vehicle and property damage to about how the information may be relevant to your company. You are provided with SUV insurance loss data from 2012-2014 from the Insurance Institute for Highway Safety Highway Loss Data Institute (IIHS-HLDI) (http://www.iihs.org/iihs/topics/insurance-loss-information). The data are provided in the Excel spreadsheet SUVInsuranceLossData.xlsx that is linked with this assignment. The data include size classification (Small, Midsize, Large), vehicle name and insurance loss scores for two vehicle and property damage categories. There is a description of how to interpret the reported scores on the IIHS-HLDI website. There is also a link on that site to a page with auto insurance basics—this will help you understand the definitions of collision and property damage. Also notice that as part of the SUV name, there is an indicator of 2WD or 4WD. For this assignment you are to conduct an exploratory data analysis (EDA) of the loss data and then write a one-page memo to your boss summarizing your findings and providing your recommendations.

Your EDA should include: appropriate comparative plots and summary statistics should be generated for the groups identified below in a) – c). You are generating these summaries for each of collision and property damage.

a. All SUVs b. SUVs separated by size class c. SUVs separated by drive class (2WD vs. 4WD).

Be sure to think about axis scaling for any comparative plots you include. Comment on the distributions you observe, as well as any outliers. Do you need to be concerned about any outliers? Also, the written EDA includes brief descriptions of what you see based on the plots and tables (there is an example/template posted on the assignment page). Beware of the correct usage of Figure and Table captions—the example shows standard usage for published journal articles. Note: we will treat these data as samples, but they are probably not samples. The IIHS- HLDI should have information for all SUV models in the population. The data values themselves represent averages over many accidents for each of the SUV types.

Your one-page memo should include: 1. A brief summary of your findings from the EDA. Make sure you are clear about whether

or not your summaries are based on means/SDs or medians/quartiles, depending on the distributions you observe for the various loss data. If you want to point to a particular Table or Figure in your memo, make sure to reference it.

2. Your recommendation to your boss about your company’s SUV insurance policies.

2

Other Constraints You only submit one report per group, limited to a total of 5 pages. Include your group number and individual group member names in the “From” portion of your memo. Your report should be organized as follows: Page 1 is your one-page memo that includes your summary and recommendations. Feel free to summarize your findings and recommendations using bullets if it helps your organization, but be sure to include introductory sentences to any list of bullets. Pages 2-5 contain the results of your EDA (2 pages per variable). Four pages of EDA output is the maximum. Be thoughtful with your presentation. The EDA will provide the supporting material for the content of your memo. You may do a bit more EDA than you include in your write-up; alternatively, you may include some EDA in your write-up that is not mentioned in the memo. Remember that EDA is your way to explore the data and decide what information is useful/relevant to your final assessment. Your EDA should start with a short (200 words max) introduction that describes the data in your own words (do not just cut and paste from the website). Also, your EDA should read like a report. Do not just include a bunch of tables and figures without any references to them. Tables must have captions at the top of the table. Figures must have captions at the bottom of the figure. Number your Tables and Figures in the order in which they appear in the report (e.g., Table 1 or Figure 2). Any table or figure must be referenced in the body of the EDA report. Always provide a brief summary of what information you takeaway from any table or figure included in your report. Note that your total report is a maximum of 5 pages. Use tables for summary statistics and figures for graphs. Create your graphs so that you can scale them appropriately into a nicely formatted report. Don’t forget to fully label any graphs. Since you are providing Figure captions, your graphs should not have titles on them. This assignment will be graded on the thoughtfulness and completeness of your EDA, your ability to summarize your EDA findings and provide recommendations, as well as overall organization, formatting and quality of written product (your memo should sound like it was written by one person). While this is a statistics class, writing matters. You need to be able to communicate any statistics you do with various audiences.