STAT 628 module2 group10 Fall2022 UW-Madison

Module 2: Body Fat Percentage Analysis

Project information

Team member:

  1. Marwa Eltgani (eltgani@wisc.edu)
  2. Sheng Huang (shaung452@wisc.edu)
  3. Samach Sathitvudh (sathitvudh@wisc.edu)
  4. Xinyan Wang (xwang2587@wisc.edu)

The purpose of this project is to perform a modern statistical analysis from the BODYFAT dataset. Our methodology includes:

  • Exploratory Data Analysis: to investigate the variables' distributions along with the outliers
  • Data cleansing: to clean the data by imputing some observations which areirrational and create a new cleaned dataset
  • Statistical Model: to construct a model that effectively infers and describes the variables based on the existing data. We also conduct the experiment to select the best model
  • Model diagnostics: to investigate the proposed model whether it is appropriate for predicting body fat using several plots and tests for model assumptions
  • Model strengths and weaknesses: to analyze the finalized model and its prediction performance. We also include the discussion for the improvement and future work
  • Shiny App: to develop a real-time web application and predict a bodyfat based on our model

Content

data

  • This contains both raw and cleaned data sets we used in the analysis.

code

  • This contains R scripts for the analysis ranging from reading data, exploratory data analysis, data cleaning, data modeling to model diagnostics as well as script for Shiny app.

image

  • This contains images obtained from the code and analysis such as figures, tables, and plots.

Executive Summary

  • The two-page document that summarizes the analysis including the introduction, data cleaning, model selection and diagnostics as well as the strengths and weaknesses of the model.

Link to Shiny App

https://bodyfatcalculator.shinyapps.io/bodyfat/