📝 HW1

Due date: Friday, September 4 at midnight.

Some components of this assignment have not been seen by a previous cohort of STAT 131A students, so there may be some unforeseen hiccups.

📮 Submission

Submit your assignment via Pensive. The Pensive portal will be live at least a few days before the HW deadline.

  • Make sure to tag your answers properly on Pensive, or else you may be docked points for making the grading process more time-consuming.

For coding components, you will produce both (1) a .qmd file with your code and (2) an PDF containing the code and output.

  • On Pensive, you will submit a single ZIP file containing both the .qmd and PDF files.
  • Proofread your PDF to make sure all of your answers and plots are visible.

For the plot presentation, you will submit a link.

  • Submit a public link to your screencast via Pensive. Test your link in an incognito window to ensure it’s public. We will dock points and/or slip days if the link is not publicly accessible during grading.

📈 Data manipulation and plotting with R (50% of the HW1 grade)

Complete the HW1 coding problems: DataHub

  • The HW1 coding problems are located in 131a-code-fall-2026/hw/hw1 on DataHub.

  • You are free to download the HW1 repository and work locally in RStudio or Positron, if you wish.

  • Be scrappy–if you don’t know how to navigate something in RStudio or Positron, ask an LLM or Google your question before posting to Ed!

  • If you do post to Ed, post publicly whenever possible so everyone can benefit from the answer.

🗣️ Plot presentation (50% of the HW1 grade)

This plot is also referenced in the code notebook linked above. You may find it helpful to read the code notebook before completing this portion.

In their 2013 research paper titled The Missing “One-Offs”: The Hidden Supply of High-Achieving, Low-Income Students, Caroline Hoxby and Chris Avery investigate the behavior of high-achieving low-income applicants to undergraduate programs in the United States.

Here is Figure 10 of their research paper:

Your task: Record a 60-90 second screencast describing the plot above.

Your audience is a UC Berkeley undergraduate who has not taken a statistics or data science course, and is unfamiliar with the research paper.

  • You do not need to read the research paper in full to understand the plot, but you might find it helpful to reference.

In your screencast, you must explain the minimum necessary background required to understand the plot, along with the key takeaway(s) of the plot.

  • You should use your mouse pointer to indicate particular points of interest on the plot. Alternatively, you can verbally direct the viewer to points of interest (e.g., “In the top right corner, you can see that…”).

  • You should use an engaging tone that sounds as though you are presenting to a live audience.

  • I discourage you from reading a script. It can be very helpful to write a script for a presentation ahead of time, even if you do not actually read the script when presenting.

  • Remember the three key guiding questions you should address before digging into the details: (1) What’s on the X axis? (2) What’s on the Y axis? (3) What does a specific point/line/feature on your plot mean in context?

There are lots of free tools for recording screencasts. For example, QuickTime is a useful tool for recording screencasts on a Mac. Feel free to post on Ed if you cannot identify a way to record a screencast anonymously.

  • Your name and face should not be in the screencast. In other words, the plot should take up the entire window of the screencast, with your voice playing in the background. Do not introduce yourself or use any identifying information.

  • If you need more than 60-90 seconds to record for accessibility reasons, please let the teaching staff know via Ed.

💾 Upload your screencast to your UC Berkeley Google Drive account, or another place that allows you to share a public link.

  • You must submit your public link. See submission details at the top of this page.

Why complete this problem? Communication is a critical, but often under-appreciated, component of the data science life cycle. This exercise helps develop your data storytelling ability, which is essential for getting anyone to actually care about your statistical analyses!