Reference no: EM133840236

Statistical Data Analysis and Visualisation

Coursework - Determining factors for Body Fat

Assessment Requirements

The assessment consists of two parts. An overall mark for your work will be based on a maximum of 90 marks for correct answers to the set questions and a maximum of 10 marks for overall presentation of the coursework. Your report will be assessed according to the assessment criteria provided in Section II.

You should complete the analysis in R and write a report using either word or LATEX. Where you use equations, you should make use of Microsoft Equation tools available within the Word processor. You should submit your file in PDF format only. You should check the pdf document before submitting to make sure everything is how you wanted it presented. Use an appropriate referencing style to cite where it is required. An example report will be provided as a guidance.

As shown in the table above, a page limit of between 5 to 7 pages is recommended. The report, together with title page, contents page, references, references and appendices must not exceed 12 pages.

You are free to select the style of the report (i.e., section headings and format, etc.) although it must obviously address the content listed in the table above.

Coursework: Determining factors for Body Fat
Body Fat dataset (BodyFat.csv) contains information on the relationship between physical attributes (density, age, weight, height, neck, chest, abdomen, hip, thigh, knee, ankle, biceps, forearm, hips) and their impact on body fat on 252 individuals ( Ref: Vanderbilt Statistics Datasets.

It can be used to study how these features influence body fat and develop predictive models for estimating it.
density: a person's overall body density through methods like underwater weighing (hydrostatic weighing).

age: a person's age (in years). weight: weight measure (in pounds). height: height measure (in inches) neck: neck measure (in cm).
chest: chest measurement (in cm). abdomen: abdomen measurement (in cm). hip: hip measurement (in cm).
thigh: thigh measurement (in cm). knee: knee measurement (in cm). ankle: ankle measurement (in cm). biceps: biceps measurement (in cm). forearm: forearm measurement (in cm). hips: hips measurement (in cm).
body fat: application of the Siri equation to estimate body fat percentage from the measured density (in percentages).

Visualise the distributions of each variable and provide their summary statistics.

How many variables are in the dataset?

What types of variables are they?

iv) Is the dataset balanced? Explain.

vi) Compute the summary statistics of the variables. Discuss.

Now, test the assumption that all predictor variables are independent using correlation coefficients or tests of associations. Visualise your results using scatter plots and a correlation matrix.

Using linear regression modelling, describe the variables which you believe have influence on body fat and describe your model with coefficients. Note, you may consider other models beyond taught material which may be more appropriate for modelling the outcome variable. Can you do my assignment for me? We sure can!

Split the body fat data into two, creating a new variable Bodyfat-split. Assuming all predictor variables are independent of one another, perform statistical tests to assess differences in central tendencies of your predictor variables with respect to the newly formed categorical variables Bodyfat-split. Visualise your results appropriately. Write your assumptions and state your hypotheses.

Using ANOVA testing or the nonparametric equivalent, Kruskal- Wallis testing, to assess differences in central tendencies of the interval predictor variables (eg height, weight etc.) with respect to body fat. For this, you may want to create three or four age groups. Perform additional post-hoc tests. Write your assumptions and state your hypotheses.

Perform a PCA Analysis and Hierarchical Model on this dataset to identify the underlying factors that contribute to body fat decomposition. Discuss your findings. You may use other models beyond taught material in the course.