Due on Sunday June 2 at 11:59 PM
Read the instructions carefully and make sure you have all the components on the checklist.
Set up
GitHub steps:
- Fork the workshop 8 repo on GitHub.
- Rename the repository to
lastname-firstname_homework-03(use your own last name and first name!).
RStudio/code steps:
- Clone your repository to your computer.
- Create a new Quarto document for your homework submission in the
codefolder.
- At the top of your document, include a link to your forked repository.
- Do all of your set up (reading in packages/data) at the top of your document.
Throughout the course of doing your homework, make at least 10 commits and pushes to the remote.
Problems
Problem 1. Multiple linear regression: model selection and construction (52 points)
Use the information from the homework-starter-doc.qmd to do this problem.
a. Make a table or list of all the models from class and the last one you constructed on your own. Write a caption for your table. (8 points)
Caption: table captions typically go above the table. Number the table and provide a title. Describe what is in the table (columns and rows).
Table: In your table, each row should be a model with the model number (1, 2, 3, etc.) and the predictors for each model.
There should be 5 models total (null model + 4 models with predictors).
b. Write a 5-6 sentence “statistical methods” section. (8 points)
Your answer should be in paragraph form and include:
- how you addressed the central question(s) (i.e. to examine the influence of ____, ____, and ____ on _____, I…)
- how you chose the final model (i.e. to determine the model that best described ____, I…)
- how you visually determined that your final model conformed to the assumptions of a linear model (i.e. to evaluate linear model assumptions, I…)
c. Make a visualization of the model predictions with underlying data for your “best” model. (20 points)
Show and annotate all your code. For full credit:
- make the underlying data more transparent than the model predictions
- display species names in full (not as species codes like ENCCAL or ESCCAL)
- display water treatment types in full (not as WW or DS)
- represent well-watered and drought stressed treatments with different colors
- use colors that are not the default
ggplot()colors
- facet your plot by species
- remove the legend
- finalize the plot
Make sure that the only output is the visualization!
d. Write a caption for your visualization. (6 points)
Include a data citation.
e. Write a 3-4 sentence results section. (10 points)
Your answer should be in paragraph form and address the following points:
- what predictors “best” described total mass (include model statistics here)?
- on average, what differences did you find between water treatments?
- on average, what differences did you find between species?
Problem 2. Affective visualization (24 points)
In this problem, you will create an affective visualization using your personal data in preparation for workshop during week 10.
In lecture, we talked about the three vertices of data visualization: 1) exploratory, 2) affective, and 3) communicative. We’ve done a lot of exploratory and communicative visualization, but have yet to think about affective visualization.
When thinking of affective visualization, you can expand your ideas of what data visualization could be. Some examples of affective visualizations include:
- Jill Pelto’s paintings
- Lorraine Woodruff-Long’s warming strips quilt
- Stefanie Posavec and Giorgia Lupi’s Dear Data project
Before starting, update your spreadsheet of observations.
a. Describe in words what an affective visualization could look like for your personal data (3-5 sentences). (2 points)
b. Create a sketch (on paper) of your idea. (2 points)
Include a photo of this sketch in your submission.
c. Make a draft of your visualization. (12 points)
Feel free to be creative with this! You do not have to do this in R. You could create a sculpture, painting, textile object, etc.
If you are making your visualization in R, show the annotated code and the output.
If you are making your visualization outside of R, include a photo of your visualization in your submission.
d. Write an artist statement. (8 points)
An artist statement gives the audience context to understand your work. Write 4-5 sentences to address:
- the content of your piece (what are you showing?)
- the influences (what did techniques/artists/etc. did you find influential in creating your work?)
- the form of your work (written code, watercolor, oil painting, etc.)
- your process (how did you create your work?)
Problem 3. Statistical critique (36 points)
At this point, you have seen and created a lot of figures for this class. Revisit the paper you chose for your critique and your homework 2, where you described figures or tables in the text. Address the following in full sentences (3-4 sentences each).
For this section of your homework, you will be evaluated on the logic, conciseness, and nuance of your critique.
a. Revisit and summarize (6 points)
What are the statistical tests the authors are using to address their main research question?
Insert the figure or table you described in Homework 2 here.
b. Visual clarity (10 points)
How clearly did the authors visually represent their statistics in figures? For example, are the x- and y-axes in a logical position? Do they show summary statistics (means and SE, for example) and/or model predictions, and if so, do they show the underlying data?
c. Aesthetic clarity (10 points)
How well did the authors handle “visual clutter”? How would you describe the the data:ink ratio?
d. Recommendations (can be longer than 4 sentences, 10 points)
What recommendations would you make to make the figure better? What would you take out, add, or change? Provide explanations/justifications for each of your recommendations.
Alternatively, if they did not represent their statistics in a figure, what kind of figure would you recommend to them? Describe the x- and y-axes, with any geometries and aesthetics (colors, etc.). Provide enough explanation such that someone would be able to create the figure you describe in code. Provide explanations/justifications for the figure you would recommend.
Checklist
Your homework should
- Include your name, the title (“Homework 3”), and the date you turned in the assignment (3 points)
- Include for Problem 1:
- a table for part a (if you made it in code, show the code and annotations)
- written response for part b
- full work (code and annotations) and output for part c
- written response for part d
- written response for part e
- Include for Problem 2:
- written response for part a
- photo for part b
- code and annotations or photo for part c
- written response for part d
- Include for Problem 3:
- written responses for part a-d
- written responses for parts g-h
- be uploaded to Canvas as a single PDF (1 point)
- be organized and readable (10 points), meaning that
- you are submitting a rendered PDF
- all messages and warnings are hidden
- all figures and tables are displayed correctly in the rendered PDF
- all code annotations are visible in the rendered PDF
- you have a set up chunk at the beginning and each section only has the code corresponding to that specific section (for example, no reading in packages/data in the chunk to create a visualization)
- you are submitting a rendered PDF
- be associated with a GitHub repository (15 points), meaning that
- you have a link to your GitHub repository at the top of your document
- you have committed/pushed at least 10 times over the course of completing your homework
- at least 10 of your commits/pushes contain a descriptive, concise commit message (a few words describing what changes you are committing)
- you have a link to your GitHub repository at the top of your document
141 points