Instructions: Homework #10

Obtaining and setting up the homework

• Obtain the homework template from your RStudio Cloud class project by running the following code in the R Console:

  library(ds4b.materials) # Load the class library
  launch_homework(10)      # Launch Homework 10

• You must set an RMarkdown theme and code syntax highlighting scheme of your choosing in the YAML front matter. These links will help you:

  • Choose your favorite theme among the pre-packaged themes (ignore everything below “Even More Themes”) shown at this link
  • Choose your favorite syntax highlighting among these options at this link

• Make sure your Rmd knits without errors before submitting. If it does not produce an HTML output, this means it does not knit. DO NOT SKIP THIS STEP! Ensuring code runs without errors is MORE IMPORTANT than writing code in the first place.

  • If there are errors in your code, you should comment out the code so that it does not actually run. This is BETTER than keeping the buggy code in there without commenting out - it shows me you attempted the code, but understood that it didn’t work properly. Partial credit will come to you! But, if you leave buggy code in, then the Rmd will not knit and there will be deductions.

• As always, you are encouraged to work together and use the class Slack to help each other out, but you must submit YOUR OWN CODE.
  
  

Questions

Caution: This homework is VERY short, which means each question is worth a lot of points. This is both a blessing and a curse. Plan accordingly.

This assignment will use an dataset of various physical measurements from 250 adult males. Your goal for this assignment is to build and evaluate a model from this data to predict body fat percentage (column Percent) in adult males, and then use this model to predict future outcomes. Age is measured in years, weight in pounds, height in inches, and all other measurements are circumference measured in cm.

1. Prepare the data for use: You’ll notice a column called id in the dataset, which is a single identifying number for each male in the study. This is not “measured data”, and we definitely don’t want to end up modeling it. Create a new version of the dataset called bodyfat_noid that does not contain the id column. This is a one-line pipeline - don’t overthink! Use the bodyfat_noid dataset for questions 2-5, but not 6.

2. Perform model selection: Using the step() function, determine the most appropriate model to explain variation in bodyfat percentage in this data, and then do ONE of the following (pick your favorite!): reveal the full model output with the function summary(), reveal the full model output with the broom function tidy() (you’ll need to load broom in the setup chunk!), OR reveal the model formula.

   After the code, write a properly-formatted (make sure to knit and check!) bullet point list stating the final predictors. You do not need to include information about the specific coefficient values or P-values - just bullet out what the predictors are. Make sure you have properly formatted bullet points in your knitted output!! The purpose of this part of the question is for you to practice writing markdown bullet points and ensure they knit properly.

3. Evaluate your model with testing and training splits: Create a testing-training split where the training data comprises a random 65% subset of the total bodyfat dataset. Build the model using the training data, and then evaluate how the model performs on both the training and testing splits. After the code, provide a brief explanation (1-2 sentences) about whether the model is likely overfit to the training data or not, and whether these results suggest the model will do a good or poor job predicting future outcomes.

   Your code must have the following components:

   • First, set your seed AS YOUR PERSONAL BANNER ID, using the variable defined in the setup chunk. YOU MUST USE YOUR BANNER ID AS THE NUMBER, which you should have defined in the setup code chunk!!! If you do not do this properly, there will be unpredictable mismatches between your written answer and the results your code produces when run. You must set your seed, and use it consistently.
   • Create a variable to represent the 65% split instead of hardcoding this value.
   • There is other code you need to run to help you interpret RMSE. This code should be included in the code chunk as part of the homework (i.e., not run separately in Console.)
   • Note that the functions rsquare() and rmse() are part of the modelr package - you will need to load that package in your setup chunk if it isn’t already loaded!

4. Some data fun: Make either a faceted barplot or a lollipop plot to simply show your \(R^2\) and RMSE results. To do this, you need to create from scratch a properly structured tibble that contains columns you want to plot! I cannot recommend enough the need to draw this out first before proceeding. If you do not draw this out to plan, you will be stuck here forever. Please, draw. Draw the data, draw the plot, just draw!!

   To achieve a plot, you will need to create a tibble with columns as follows (get_help("tibble")). The tibble should end up having four rows and three columns called:

   • dataset, which contains values “testing” and “training”
   • metric, which contains values “RMSE” and “R2” (don’t worry about superscript! “R2” is fine.)
   • value, which contains the relevant values for each metric

   Then, you should make a plot showing the value for each dataset, where the plot is faceted by metric and uses separate axis scales for each facet (see the faceting notes from 10/6/21 on the class website!). You should also add a color or fill (depends on what type of plot you make) mapped to dataset. The figure should be cleanly and professionally visible in the knitted Rmd, but you do not need to export it to a file with ggsave().

5. Prediction: A new man has arrived, and we want to use our model (as built with the training data split!) to predict his body fat percentage. Predict what the body fat percentage will be for a man with the following physical attributes. Your code should reveal the predicted body fat percentage.

   Hint: Do you need all this information? In fact you do not! For full credit, make sure your code is only considering body measurements that are relevant for answering the question. In other words, your code should not use any data the model doesn’t need.

   • Age: 38
   • Weight: 175.2
   • Height: 62.25
   • Neck: 31.4
   • Chest: 107.4
   • Abdomen: 93.3
   • Hip: 114.7
   • Thigh: 68.5
   • Knee: 34.1
   • Ankle: 26.2
   • Biceps: 32.2
   • Forearm: 35.4
   • Wrist: 15.9

6. Plotting: Wrangle the original bodyfat (not bodyfat_noid) dataset to instead have four columns (id, Percent, measurement, value) as a longer version of the dataset. The measurement column should be a character column of the different bodily measurements, and value should contain the corresponding numeric values. Then, use this dataset to make a plot showing the relationship between Percent and each of the other measurements, specifically only those that were predictors in the model (this means at some point in your wrangling, some data has to be removed). This figure should be faceted scatterplot where you are plotting Percent against value and faceting by measurement. Each panel should contain its own line of best fit, and you may wish to use free panel scales. Display the figure legibly in the knitted Rmd, but you do NOT need to export to a file with ggsave(). Style the figure as you like as long as labeling is clean! (Hint: This is wildly similar to the last question on HW9!!).